CPU Fan Fail warning on ASUS A7N8X Deluxe but all seems OK.

A

Alan Walker

Hi,

I've got one of these with an XP3000, bought it pre-built with CPU and
memory. Every time it starts up I get a 'CPU Fan Fail' warning.

POST completes OK and machine boots. No signs of actual overheating.

Sometimes if I do a restart rather than a cold start I don't get the
message.

Is this just the fan taking time to spin up ?

Fan is spinning, BIOS and probe show it running steadily at 2100 - 2200RPM.

Cpu temp idles at around 51C and if I hammer the machine heavily it rises to
about 66C, dropping fairly rapidly back to idle temperature. It never hits
the alarm point I've set at 72C.

Should I be worried or just mildly annoyed ?

All comments appreciated.

Alan
 
R

Rob Hemmings

Alan Walker said:
Hi,

I've got one of these with an XP3000, bought it pre-built with CPU and
memory. Every time it starts up I get a 'CPU Fan Fail' warning.

POST completes OK and machine boots. No signs of actual overheating.

Sometimes if I do a restart rather than a cold start I don't get the
message.

Is this just the fan taking time to spin up ?

Fan is spinning, BIOS and probe show it running steadily at 2100 - 2200RPM.

Cpu temp idles at around 51C and if I hammer the machine heavily it rises to
about 66C, dropping fairly rapidly back to idle temperature. It never hits
the alarm point I've set at 72C.

Should I be worried or just mildly annoyed ?

All comments appreciated.
Click to expand...

Same mobo.
I get a similar result if I turn the Aero7 HSF controller knob
down to minimum as it then runs below 600rpm, which ISTR
is the alarm point. On next cold boot, I get the same message
as you, so it sounds as if your CPU fan is just a bit slow to
start up - have you tried physically watching it from a cold boot?
If you have a retail AMD HSF, it's easy to change the fan.
HTH
 
A

Alan Walker

Rob said:
Same mobo.
I get a similar result if I turn the Aero7 HSF controller knob
down to minimum as it then runs below 600rpm, which ISTR
is the alarm point. On next cold boot, I get the same message
as you, so it sounds as if your CPU fan is just a bit slow to
start up - have you tried physically watching it from a cold boot?
If you have a retail AMD HSF, it's easy to change the fan.
HTH
Click to expand...

Thanks, been building/fixing machines for years but this is my first that
talks to me during POST.

I've watched it and seems to be running OK. It's also fairly quiet so no
obvious bearing problems.

Alan
 
R

Rob Hemmings

Alan Walker said:
Thanks, been building/fixing machines for years but this is my first that
talks to me during POST.

I've watched it and seems to be running OK. It's also fairly quiet so no
obvious bearing problems.
Click to expand...

I suppose as long as your temps are ok, you could just turn
the spoken report off in the BIOS - not a real fix I know,
but as the mobo will auto-shutdown if the CPU temp goes too
high (say the fan fails completely), not really a dangerous thing
to do (I must admit I got fed-up with it, especially if I'd left
the volume control on the speakers set high when I powered-
off - being woken by 'her' at 90dB at 7am was a bit too much
for me after a night on the beer! :)
 
T

the gnome

What RPM is the fan supposed to run at, my CPU fans run about twice that
speed. I did have one that had a faulty cable and sometimes it registered
half speed even though the fan was running without change.

Can you change the RPM alarm level in BIOS, it may be set too high for this
fan.

the_gnome
 
K

kony

Hi,

I've got one of these with an XP3000, bought it pre-built with CPU and
memory. Every time it starts up I get a 'CPU Fan Fail' warning.

POST completes OK and machine boots. No signs of actual overheating.

Sometimes if I do a restart rather than a cold start I don't get the
message.

Is this just the fan taking time to spin up ?

Fan is spinning, BIOS and probe show it running steadily at 2100 - 2200RPM.

Cpu temp idles at around 51C and if I hammer the machine heavily it rises to
about 66C, dropping fairly rapidly back to idle temperature. It never hits
the alarm point I've set at 72C.

Should I be worried or just mildly annoyed ?

All comments appreciated.

Alan
Click to expand...

Many boards will issue such a warning when the CPU fan rpm is below a
certain threshold. There might be a bios update that addresses this, but
it might be more significant that your temps are up around 66C. Run a
program like "CPUBurn" (Google search) for 30 minutes to check stability.
The cooling system should be confirmed adequate for an extended full load
condition.
 
A

Alan Walker

kony said:
Many boards will issue such a warning when the CPU fan rpm is below a
certain threshold. There might be a bios update that addresses this,
but it might be more significant that your temps are up around 66C.
Run a program like "CPUBurn" (Google search) for 30 minutes to check
stability. The cooling system should be confirmed adequate for an
extended full load condition.
Click to expand...

It's now idling at 50C, hits 66C if I play hi-res 3D games for a long time,
is this worryingly high ?
 
A

Andy

Alan Walker said:
It's now idling at 50C, hits 66C if I play hi-res 3D games for a long time,
is this worryingly high ?
Click to expand...

Does it crash ? If not no it is fine. If it does crash yes it is a bit
high.

Hope this helps

Andy
 
K

kony

It's now idling at 50C, hits 66C if I play hi-res 3D games for a long time,
is this worryingly high ?
Click to expand...


It needs to be adequate for FULL LOAD condition. Not some certain game
but with the CPU at 100% utilization for an extended period of time. That
is a very basic requirement that 100% of all systems should meet. If you
haven't tested that yet, you don't yet know what the max temp is and so
can't yet determine if the cooling is adequate, if the system would be
stable. This isn't even considering summer... if summer will bring higher
ambient temps then you can expect the CPU temp to rise a corresponding
amount.

Many games do not put the CPU at full load. We could try to guess which
ones would but there's no point in that when a simple program like CPUBurn
will do all the work.
 
A

Alan Walker

kony said:
It needs to be adequate for FULL LOAD condition. Not some certain
game but with the CPU at 100% utilization for an extended period of
time. That is a very basic requirement that 100% of all systems
should meet. If you haven't tested that yet, you don't yet know
what the max temp is and so can't yet determine if the cooling is
adequate, if the system would be stable. This isn't even considering
summer... if summer will bring higher ambient temps then you can
expect the CPU temp to rise a corresponding amount.

Many games do not put the CPU at full load. We could try to guess
which ones would but there's no point in that when a simple program
like CPUBurn will do all the work.
Click to expand...

Things get stranger, from idle at 50C I ran CPUBurn, temp rose over ten
minutes to 57C be levelled off and after 30 minutes burn still at 57C. Next
thing that occurs to me might be GPU and CPU running together at full
throttle might pump too much heat into the case, got a dual-fan PSU moving
quite a bit of air through and one case fan, which I suspect is not at it's
best. Maybe next step is to put an exhaust fan in.
 
K

kony

Things get stranger, from idle at 50C I ran CPUBurn, temp rose over ten
minutes to 57C be levelled off and after 30 minutes burn still at 57C. Next
thing that occurs to me might be GPU and CPU running together at full
throttle might pump too much heat into the case, got a dual-fan PSU moving
quite a bit of air through and one case fan, which I suspect is not at it's
best. Maybe next step is to put an exhaust fan in.
Click to expand...


That is an interesting result.

Just to confirm, you did set CPUBurn to "K7" setting and "High" priority?

Leaving it like that for 30 minutes should show quite close to peak temps
(per ambient temp, only you know if the room temp changes over time).

Considering the above, and your results, it does seem as though you might
have insufficient case airflow. If your power supply and/or case fan have
stamped-in-metal grills you ought to cut out those grills, put in chrome
wire grills if you still need some protection from the blades.

Generally the best cooling is attained from a rear case wall fan, right
beneath the power supply, blowing out, and the majority of the case air
intake vents on the bottom front. Two rear fans (including the power
supply rear fan) should be sufficient but a system optimized for
low-noise, therefore low RPM fans, may need three exhaust or to use > 80mm
fans with high-performance (heat) parts inside.
 
A

Alan Walker

DaveW said:
Your CPU is running rather hot at those temps. You would like idle
to be around 37C.
Click to expand...

Is an extra 10C going to do any harm until I get a new case (couple of
weeks) ?
 
G

Guest

the gnome said:
What RPM is the fan supposed to run at, my CPU fans run about twice that
speed. I did have one that had a faulty cable and sometimes it registered
half speed even though the fan was running without change.

Can you change the RPM alarm level in BIOS, it may be set too high for this
fan.
Click to expand...

I didn't explain that very well - I *really* don't want to be woken-up
by a system that's telling me it's ok! That's the problem if you leave
the (by default) spoken BIOS report set to 'on' (doesn't even need
a working CPU or RAM to operate - it's self-contained 'chip', audio
routed to line-out socket on backplane).
'Tis nice to have the option though. XLNT mobo, and the Aero 7
series of fans (all sorts) are good vfm.
HTH
 
A

Alan Walker

Alan said:
Is an extra 10C going to do any harm until I get a new case (couple of
weeks) ?
Click to expand...

I give up trying to work this out, it's dropping about 1C in idling
temperature every day.

I've bought a full tower case on eBay and I'm going to rebuild in that with
a handful of big fans.
 
K

kony

No, so long as the temp is low enough that there are no errors, to data
for example.

I give up trying to work this out, it's dropping about 1C in idling
temperature every day.
Click to expand...

A freshly installed CPU may do that for a few days, particularly when
there was either

A) Too much heatsink compound
B) OEM thermal interface material
C) Phase-change type compound like synthetic Arctic
Silver/Ceramique/Alumina/etc.
I've bought a full tower case on eBay and I'm going to rebuild in that with
a handful of big fans.
Click to expand...

Don't overdo it or you'd curse the noise levels... generally very low RPM
(~2000RPM) fans are adequate in larger sizes or when using more than
one, providing they're strategically located.
 
A

Alan Walker

kony said:
No, so long as the temp is low enough that there are no errors, to
data for example.



A freshly installed CPU may do that for a few days, particularly when
there was either

A) Too much heatsink compound
B) OEM thermal interface material
C) Phase-change type compound like synthetic Arctic
Silver/Ceramique/Alumina/etc.
Click to expand...

Ah !

OK, thanks.
Don't overdo it or you'd curse the noise levels... generally very low
RPM (~2000RPM) fans are adequate in larger sizes or when using more
than one, providing they're strategically located.
Click to expand...

The idea was to get a few big slow ones and run them with minimum airflow
impedance, i.e. cut all the grills out of the case. No kids to insert
fingers so safety's not a problem.
 
P

Parish

Alan said:
Is an extra 10C going to do any harm until I get a new case (couple of
weeks) ?
Click to expand...

I shouldn't think so. Athlons run hotter than Pentiums anyway. The
maximum _operating_ temperature of my (Barton-cored) Athlon MP2800s is
90C (this varies with the core type and voltage, but they're all in the
same region). Mine idle in the high 40's and rise to the low 50's under
load.

BTW, Athlons seems pretty robust. I once left my PC for 10 minutes, came
back and found it had shutdown, WTF?, powered it back up and got a BIOS
mesage about fan failure or temperature (can't remember the exact
message), went into the BIOS, CPU1 was running at......104.8C!!!! Sh*t!,
powered down, opened the case, nasty burning smell.

The cause? The wiring loom to the ATX m/b connectors was clipped to the
PSU (to keep it out of the airflow) using a tie-wrap through a self
adhesive "bracket" that was supplied with the case. Obviously not a very
good adhesive as it had unstuck itself and the loom dropped down. It
wouldn't have been a problem except that I'd forgotten to cut the tail
off the tie-wrap and it had stuck in the fan blades stalling the fan.

The CPU still works perfectly, much to my amazement but, when I searched
AMD's ite for thermal information I found a doc relating to the design
of thermal protection systems which states that the maximum threshold
temp. for shutting down the system should be 125C!!! (and it must
shutdown within 500ms).

Regards,

Parish
 
Top