lm_sensors reports 3.3V too high

lm_sensors reports 3.3V too high

Post by waxinwax.. » Sat, 17 Jun 2006 17:52:32



I recently moved and now experience occassional (non reproducible)
internal compiler errors with gcc. I never had problems with this
machine at my old address. I see that lm_sensors reports that certain
voltages are too high:

as99127f-i2c-0-2d
Adapter: SMBus I801 adapter at e800
VCore 1:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
VCore 2:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
+3.3V:     +3.52 V  (min =  +3.14 V, max =  +3.46 V)       ALARM
+5V:       +5.00 V  (min =  +4.73 V, max =  +5.24 V)
+12V:     +11.55 V  (min = +10.82 V, max = +13.19 V)
-12V:     -11.89 V  (min = -13.22 V, max = -10.74 V)
-5V:       -5.03 V  (min =  -5.25 V, max =  -4.74 V)
fan1:     4326 RPM  (min =    0 RPM, div = 2)
fan2:        0 RPM  (min =    0 RPM, div = 2)
fan3:        0 RPM  (min =    0 RPM, div = 2)
M/B Temp:    +33C  (high =  +105C, hyst =    +0C)
CPU Temp:  +20.5C  (high =  +100C, hyst =   +92C)
temp3:     -31.5C  (high =  +122C, hyst =  +121C)
vid:      +1.750 V  (VRM Version 8.2)
alarms:
beep_enable:
          Sound alarm enabled

VCore 1 & 2 were previously too high as well (both > 1.82V & < 1.83V).
I've checked the BIOS and the VCore setting is set to it's minimum
(1.75V). Most of the BIOS reported values correspond with lm_sensors,
except for the CPU temperature which the BIOS reports as 46 C. But the
3.3V seems too high. The BIOS value is the same as lm_sensors, although
the BIOS doesn't include a high water mark, just the value.

I can't fiddle with the 3.3V setting in the BIOS, so I now wonder what
to do? The building where the machine resides is an old building from
the 19th century, so could this have to do with the electrical setup of
the building?

Thanks

 
 
 

lm_sensors reports 3.3V too high

Post by General Schvantzkop » Sat, 17 Jun 2006 21:05:04



> I recently moved and now experience occassional (non reproducible)
> internal compiler errors with gcc. I never had problems with this
> machine at my old address. I see that lm_sensors reports that certain
> voltages are too high:

> as99127f-i2c-0-2d
> Adapter: SMBus I801 adapter at e800
> VCore 1:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> VCore 2:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> +3.3V:     +3.52 V  (min =  +3.14 V, max =  +3.46 V)       ALARM
> +5V:       +5.00 V  (min =  +4.73 V, max =  +5.24 V)
> +12V:     +11.55 V  (min = +10.82 V, max = +13.19 V)
> -12V:     -11.89 V  (min = -13.22 V, max = -10.74 V)
> -5V:       -5.03 V  (min =  -5.25 V, max =  -4.74 V)
> fan1:     4326 RPM  (min =    0 RPM, div = 2)
> fan2:        0 RPM  (min =    0 RPM, div = 2)
> fan3:        0 RPM  (min =    0 RPM, div = 2)
> M/B Temp:    +33C  (high =  +105C, hyst =    +0C)
> CPU Temp:  +20.5C  (high =  +100C, hyst =   +92C)
> temp3:     -31.5C  (high =  +122C, hyst =  +121C)
> vid:      +1.750 V  (VRM Version 8.2)
> alarms:
> beep_enable:
>           Sound alarm enabled

> VCore 1 & 2 were previously too high as well (both > 1.82V & < 1.83V).
> I've checked the BIOS and the VCore setting is set to it's minimum
> (1.75V). Most of the BIOS reported values correspond with lm_sensors,
> except for the CPU temperature which the BIOS reports as 46 C. But the
> 3.3V seems too high. The BIOS value is the same as lm_sensors, although
> the BIOS doesn't include a high water mark, just the value.

> I can't fiddle with the 3.3V setting in the BIOS, so I now wonder what
> to do? The building where the machine resides is an old building from
> the 19th century, so could this have to do with the electrical setup of
> the building?

> Thanks

Power supplies are designed to handle a wide range of AC voltages so it's
doubtful if the building has anything to do with your problems. Big spikes
could be a problem so I suppose inadequate wiring combined with a lot of
noisy motors from refrigerators and elevators might cause you trouble but
not the kind of problem that you are seeing.

Does your power supply have any adjustment knobs or screws on it? If it
does then one of the adjustment screws might have turned during the move.
If you can do a manual adjustment then that should be the first thing you
do. If you can't you might need a new power supply.

 
 
 

lm_sensors reports 3.3V too high

Post by waxinwax.. » Sun, 18 Jun 2006 02:56:59




> > I recently moved and now experience occassional (non reproducible)
> > internal compiler errors with gcc. I never had problems with this
> > machine at my old address. I see that lm_sensors reports that certain
> > voltages are too high:

> > as99127f-i2c-0-2d
> > Adapter: SMBus I801 adapter at e800
> > VCore 1:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> > VCore 2:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> > +3.3V:     +3.52 V  (min =  +3.14 V, max =  +3.46 V)       ALARM
> > +5V:       +5.00 V  (min =  +4.73 V, max =  +5.24 V)
> > +12V:     +11.55 V  (min = +10.82 V, max = +13.19 V)
> > -12V:     -11.89 V  (min = -13.22 V, max = -10.74 V)
> > -5V:       -5.03 V  (min =  -5.25 V, max =  -4.74 V)
> > fan1:     4326 RPM  (min =    0 RPM, div = 2)
> > fan2:        0 RPM  (min =    0 RPM, div = 2)
> > fan3:        0 RPM  (min =    0 RPM, div = 2)
> > M/B Temp:    +33C  (high =  +105C, hyst =    +0C)
> > CPU Temp:  +20.5C  (high =  +100C, hyst =   +92C)
> > temp3:     -31.5C  (high =  +122C, hyst =  +121C)
> > vid:      +1.750 V  (VRM Version 8.2)
> > alarms:
> > beep_enable:
> >           Sound alarm enabled

> > VCore 1 & 2 were previously too high as well (both > 1.82V & < 1.83V).
> > I've checked the BIOS and the VCore setting is set to it's minimum
> > (1.75V). Most of the BIOS reported values correspond with lm_sensors,
> > except for the CPU temperature which the BIOS reports as 46 C. But the
> > 3.3V seems too high. The BIOS value is the same as lm_sensors, although
> > the BIOS doesn't include a high water mark, just the value.

> > I can't fiddle with the 3.3V setting in the BIOS, so I now wonder what
> > to do? The building where the machine resides is an old building from
> > the 19th century, so could this have to do with the electrical setup of
> > the building?

> > Thanks

> Power supplies are designed to handle a wide range of AC voltages so it's
> doubtful if the building has anything to do with your problems. Big spikes
> could be a problem so I suppose inadequate wiring combined with a lot of
> noisy motors from refrigerators and elevators might cause you trouble but
> not the kind of problem that you are seeing.

> Does your power supply have any adjustment knobs or screws on it? If it
> does then one of the adjustment screws might have turned during the move.
> If you can do a manual adjustment then that should be the first thing you
> do. If you can't you might need a new power supply.

I checked the back (visible from outside) and the surfaces of the power
supply on the inside, but there's no adjustments available. I know some
supplies have a 110-220 switch, but this one apparently has nothing
adjustable (except for on/off).

My old address had grounded electrical sockets, while this one doesn't.
I pulled a (grounded) extension cable from a grounded kitchen socket
but there's no difference with grounded sockets.

I do notice occassional flickering of my (CRT) screen, which I never
experienced in my old house (5 years of operation with this machine &
monitor). My old house used to be habited by an electrician before me.
He had redone most of the wiring when I came to live there.

I even lowered the clock settings for the memory (from 100 to 66), but
the voltage remains the same. Now the BIOS doesn't report the voltage
as an error, likely because it lacks any kind of error reporting. I'm
assuming lm_sensors' high-water/voltage mark (+3.46V) is correct, since
I really can't think of any other reason why I now get (non
reproducible) internal compiler errors.

All this inspecting has shown me something peculiar. /proc/cpuinfo says
I have an 817.416Mhz cpu, while my BIOS is set to 800Mhz. The BIOS has
a 816Mhz option which I'm going to select now. (Never realised I was an
underclocker till now).

Thanks

 
 
 

lm_sensors reports 3.3V too high

Post by waxinwax.. » Sun, 18 Jun 2006 03:13:33





> > > I recently moved and now experience occassional (non reproducible)
> > > internal compiler errors with gcc. I never had problems with this
> > > machine at my old address. I see that lm_sensors reports that certain
> > > voltages are too high:

> > > as99127f-i2c-0-2d
> > > Adapter: SMBus I801 adapter at e800
> > > VCore 1:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> > > VCore 2:   +1.82 V  (min =  +1.66 V, max =  +1.82 V)
> > > +3.3V:     +3.52 V  (min =  +3.14 V, max =  +3.46 V)       ALARM
> > > +5V:       +5.00 V  (min =  +4.73 V, max =  +5.24 V)
> > > +12V:     +11.55 V  (min = +10.82 V, max = +13.19 V)
> > > -12V:     -11.89 V  (min = -13.22 V, max = -10.74 V)
> > > -5V:       -5.03 V  (min =  -5.25 V, max =  -4.74 V)
> > > fan1:     4326 RPM  (min =    0 RPM, div = 2)
> > > fan2:        0 RPM  (min =    0 RPM, div = 2)
> > > fan3:        0 RPM  (min =    0 RPM, div = 2)
> > > M/B Temp:    +33C  (high =  +105C, hyst =    +0C)
> > > CPU Temp:  +20.5C  (high =  +100C, hyst =   +92C)
> > > temp3:     -31.5C  (high =  +122C, hyst =  +121C)
> > > vid:      +1.750 V  (VRM Version 8.2)
> > > alarms:
> > > beep_enable:
> > >           Sound alarm enabled

> > > VCore 1 & 2 were previously too high as well (both > 1.82V & < 1.83V).
> > > I've checked the BIOS and the VCore setting is set to it's minimum
> > > (1.75V). Most of the BIOS reported values correspond with lm_sensors,
> > > except for the CPU temperature which the BIOS reports as 46 C. But the
> > > 3.3V seems too high. The BIOS value is the same as lm_sensors, although
> > > the BIOS doesn't include a high water mark, just the value.

> > > I can't fiddle with the 3.3V setting in the BIOS, so I now wonder what
> > > to do? The building where the machine resides is an old building from
> > > the 19th century, so could this have to do with the electrical setup of
> > > the building?

> > > Thanks

> > Power supplies are designed to handle a wide range of AC voltages so it's
> > doubtful if the building has anything to do with your problems. Big spikes
> > could be a problem so I suppose inadequate wiring combined with a lot of
> > noisy motors from refrigerators and elevators might cause you trouble but
> > not the kind of problem that you are seeing.

> > Does your power supply have any adjustment knobs or screws on it? If it
> > does then one of the adjustment screws might have turned during the move.
> > If you can do a manual adjustment then that should be the first thing you
> > do. If you can't you might need a new power supply.

> I checked the back (visible from outside) and the surfaces of the power
> supply on the inside, but there's no adjustments available. I know some
> supplies have a 110-220 switch, but this one apparently has nothing
> adjustable (except for on/off).

> My old address had grounded electrical sockets, while this one doesn't.
> I pulled a (grounded) extension cable from a grounded kitchen socket
> but there's no difference with grounded sockets.

> I do notice occassional flickering of my (CRT) screen, which I never
> experienced in my old house (5 years of operation with this machine &
> monitor). My old house used to be habited by an electrician before me.
> He had redone most of the wiring when I came to live there.

> I even lowered the clock settings for the memory (from 100 to 66), but
> the voltage remains the same. Now the BIOS doesn't report the voltage
> as an error, likely because it lacks any kind of error reporting. I'm
> assuming lm_sensors' high-water/voltage mark (+3.46V) is correct, since
> I really can't think of any other reason why I now get (non
> reproducible) internal compiler errors.

> All this inspecting has shown me something peculiar. /proc/cpuinfo says
> I have an 817.416Mhz cpu, while my BIOS is set to 800Mhz. The BIOS has
> a 816Mhz option which I'm going to select now. (Never realised I was an
> underclocker till now).

Changing the clock for the CPU (from 800 to 816Mhz) has no influence on
the voltage. lm_sensors still reports +3.3V as too high (currently
+3.52V). Too bad I don't have lm_sensors output when the system was at
my previous address. Maybe the voltage was too high all the time, but
it's strange I only now get the ICEs. Does RAM memory degrade over
time?

Thanks

 
 
 

lm_sensors reports 3.3V too high

Post by john » Wed, 21 Jun 2006 07:17:47


The clue for me is htat you moved the box!

shut down.  Reseat every memory stick, every PCI card, and the AGP card,
if there is one.

check for a loose*rolling around that could be shorting something!

I had a similar problem, and finally, after all the above, simply
removed the PSU, and, shook it, finding that a disc capacitor was
rattling around, and could have introduced it's value across a circuit,
possibly causing 'noise' in one of the regulators.  who knows?  $19.99
for a new PSU, locally, and she is working wonderfully!

Another system had a dragging fan, due to dust, and, when cleaned, the
voltages stabilized in the safe zones!  I suppose the dragging fan
caused voltages to change, due to the overheating of hte fan and the
ripple effect upn the other switching power supply voltages.

Also, make sure all the power plugs are inserted firmly, as I had an
intermittant board, voltages that fluctuated, due to the mainboard power
connector being a bit crooked, all cleared up upon pressing it home!

 
 
 

1. lm_sensors inaccurately reporting sudden changes?

hi all,
I have lm_sensors running on a Poweredge 350.  It runs hourly as a cron
job.  Recently I had a really massive change between reports, wondered
if anyone could shed some light.

Here's the last 'normal' report:

adm9240-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 0440
2.5V:      +1.48 V  (min =  +0.00 V, max =  +1.60 V)
Vccp1:     +1.77 V  (min =  +0.00 V, max =  +1.82 V)
3.3V:      +3.28 V  (min =  +0.00 V, max =  +3.55 V)
5V:        +5.13 V  (min =  +0.00 V, max =  +5.33 V)
12V:      +11.62 V  (min =  +0.00 V, max = +12.81 V)
Vccp2:     +2.48 V  (min =  +1.39 V, max =  +2.64 V)
fan1:     8132 RPM  (min = 3125 RPM, div = 2)
fan2:     8333 RPM  (min = 3125 RPM, div = 2)
temp:      +37.5 C  (high =   +60 C, hyst =    +0 C)
vid:       +1.75 V
alarms:

And here's the 'wacky' report one hour later:

adm9240-i2c-0-2d
Adapter: SMBus PIIX4 adapter at 0440
2.5V:      +1.48 V  (min =  +0.00 V, max =  +1.60 V)   ALARM
Vccp1:     +1.75 V  (min =  +0.00 V, max =  +1.82 V)   ALARM
3.3V:      +3.28 V  (min =  +0.00 V, max =  +3.55 V)   ALARM
5V:        +5.10 V  (min =  +0.00 V, max =  +5.33 V)   ALARM
12V:      +11.62 V  (min =  +0.00 V, max = +12.81 V)
Vccp2:     +2.48 V  (min =  +1.39 V, max =  +2.64 V)   ALARM
fan1:     2033 RPM  (min =  781 RPM, div = 8)          ALARM
fan2:     2109 RPM  (min =    0 RPM, div = 8)          ALARM
temp:       -0.5 C  (high =    -1 C, hyst =    -1 C)   ALARM
vid:       +0.00 V
alarms:   Chassis intrusion detection                  ALARM

The box is located in a dedicated colo facility, so I'd be surprised if
the temp in the place dropped almost 40 degrees celsius in an
hour...and the chassis intrusion detection alert doesn't make much
sense either.  And why would the 'div' factor change on the fans?

Any insights appreciated.  thanks in advance.

2. Problems with ATAPI-CDROM and FreeBSD

3. high level performance reports

4. File System Errors -- Help!

5. REPORT: NFS and ReiserFS very high latencies, with low-latency and preempt applied

6. Linux META-FAQ

7. bug (trouble?) report on high mem support

8. Motd prior to Login/Ftp

9. high hard drive temperatures reported by smartd

10. REPORT: NFS and ReiserFS very high latencies, with low-latency and preempt applied

11. high sys% reported by sar.

12. bug (trouble?) report on high mem support

13. How High Is High ? (CPU Load)