CPU Temperature

In article ,
says...
On Thu, 29 Jul 2004 21:36:50 GMT, "DaveW" wrote:

The maximum temp while under load for that CPU should be around 50C. (100C
is the boiling point of water!)

Quite subjective while you saying that.
You have to take into account the room temperature.


For what, the boiling point of water? (just kidding)

Prescotts run a bit hotter too, so you might see temps
of 60C (assume 25C ambient temp) and yet the CPU will
still be stable. My personal comfort zone is 50-55C for
AMD CPUs. Anything over 55C and I'll consider adding a
better heatsink.

The only thing normal about those temps is people bringing back their
systems to the manufacturer with toasted CPUs.

Tell them that you'll either leave with a refund for the system or the system
if they can cut the temps in half for idle. I trust they use thermal paste
between the CPU and heatsink and I trust they have Some Kind of fan
on the heatsink. In any case, that's not your problem, it's Theirs!

Intel CPUs will throttle back when the temps reach a certain point in a
desparate attempt to lower the operating temps. The higher the temps,
the more throttling and the slower the system will run.

Your manufacturer, for these temps, is actually claiming that their systems
run slower than any other Intel 3.06. I wouldn't like to even guess just how
slow your system runs when booted to winders and running idle.

It's highly unlikely that you run any sort of higher graphics games on the
system. If you did, I suspect you could count individual frames by eyeball.

On Thu, 29 Jul 2004 21:29:44 +0100, "Nel" wrote:

I'm certain that my CPU (P4 3.06gHz) is running far too hot and is the cause
of my unstable PC. The manufacturer claims the temp is normal for this type
of system - which is more like a laptop than a desktop, as the motherboard
and cards are built in behind the TFT screen and the hardrive is built into
it's base.

According to Sisoft Sandra, the CPU runs at 108-113'C during normal
conditions an the top of the unit is hot to the touch. Can I rely on this
software's reading?

I've also heard that a P4 powers down when it reaches are certain temp to
prevent burn out - is there any software that can prove that this is
happening?

All help will be appreciated as this system has been in for repair 3 times
in 6 months and I'm thoroughly sick of it, and I want to look into the
possibility of getting a refund if they cannot supply me with a stable
system.

Thanks in advance
~~~~~~
Bait for spammers:
root@localhost
postmaster@localhost
admin@localhost
abuse@localhost
]

~~~~~~
Remove "spamless" to email me.

BTW, I did see idle temps up over 75C once, after rebuilding my system.
Turns out I didn't have the HS properly seated on the CPU. Quickly turned
off the computer, reseated the HS, and get idle temps in the 40 range now.
That's on my desktop, with a P4 2.4GHz processor.

Clint

"Nel" wrote in message
...
I'm certain that my CPU (P4 3.06gHz) is running far too hot and is the
cause
of my unstable PC. The manufacturer claims the temp is normal for this
type
of system - which is more like a laptop than a desktop, as the motherboard
and cards are built in behind the TFT screen and the hardrive is built
into
it's base.

According to Sisoft Sandra, the CPU runs at 108-113'C during normal
conditions an the top of the unit is hot to the touch. Can I rely on this
software's reading?

I've also heard that a P4 powers down when it reaches are certain temp to
prevent burn out - is there any software that can prove that this is
happening?

All help will be appreciated as this system has been in for repair 3 times
in 6 months and I'm thoroughly sick of it, and I want to look into the
possibility of getting a refund if they cannot supply me with a stable
system.

Thanks in advance

"Clint" wrote in message
news:uFtOc.140176$od7.27629@pd7tw3no...
Just to give you one more data point... My laptop (a Dell Inspiron 9100)
uses a desktop P4 CPU (3.0GHz, 800MHz bus). I haven't seen the temp over
65C.

That's about right:
o Laptops are designed to operate at higher temperatures
---- 89oC being a common absolute limit, 80oC the more usual
o That also means laptops have a lower ambient operating temperature
---- higher operating is due to the low air circulation
---- thus typically they require ambient to be below 35oC, desktop 40oC

CPU temperature wise that is ok - well within the thermal design spec.

The original poster, however, is claiming over 110oC:
o The P4 will throttle severely at those temperatures
---- as commented, the machine will operate somewhat slowly
o Additionally, it is well above the thermal design spec of ~80oC
---- thus there is a risk of damage to the mainboard (socket at least)

So...
o Try to verify that temperature - carefully
---- if you can touch the heatsink, it would be *extremely* hot
---- exit air temperature would also be *extemely* hot
---- 69oC is the touchable temperature limit before pain
o If the mainboard has onboard temperature monitoring, use it
---- Sandra may be miss-reading the temperature
o Then verify the heatsink is seated properly
---- it is quite possible to run a P4 without a heatsink (not advised)
---- in doing so it will run very hot, and very very slowly

In SFF or custom-enclosure environments, it isn't impossible for the
heatsink to be canted at one side - thus thermally somewhat ineffective.

The finger-touch is a good one to verify reported temperatures.

There is one other possibility - the temp monitoring is out of spec:
o Most thermistors have a tolerance of +/- 10%
---- so a 50oC actual temperature may be reported as ~45-55oC
o It isn't inconceivably that a tolerance of +/- 20% may apply
---- at least at the tail end of production distributions
---- since this is not a critical component re operation & QA specs
---- which would mean 50oC actual is reported as ~40-60oC
o For temps of 110oC to be reported, something is "wrong"
---- hence the need for investigation as detailed above

The least likely possibility is the reporting is in Farenheit - 113-F,
however discount that. You need to touch the heatsink & verify
it is properly seated - mechanically this is an important check too.
--
Dorothy Bradbury
www.stores.ebay.co.uk/panaflofan for quiet Panaflo fans & other items
http://homepage.ntlworld.com/dorothy...ry/panaflo.htm (Free Delivery)

Threshold of pain induced by heat is 45 degrees C; "69oC is the touchable
temperature limit before pain" is way too high.

--
Phil Weldon, pweldonatmindjumpdotcom
For communication,
replace "at" with the 'at sign'
replace "mindjump" with "mindspring."
replace "dot" with "."


"Dorothy Bradbury" wrote in message
...

Threshold of pain induced by heat is 45 degrees C; "69oC is the
touchable temperature limit before pain" is way too high.

Yes, that is the Psychophysics definition of the threshold of pain,
although figure varies according to usage, material & subject bias :-)
o Clinical Usage (eg, laser treatment):
---- Psychophysics list temperature threshold for pain as 45oC
o Design Material (eg, product liability):
---- Aluminium 45oC, Phenolic fan blade 61oC, Plastic socket 64-77oC
o Subject Bias (eg, conditioning or reward)
---- Temp perception of "pain" can be overriden beyond burning threshold
---- Coffee tasters report preference for hot coffee, despite it being 60oC
---- Yet 60oC is above the burn threshold, never mind above oral pain threshold

Ok, back to the original poster:
o Poster reports "According to Sisoft Sandra, the CPU runs at 108-113'C"

Want a binary test for a) Sandra is wrong, or b) Heatsink installed wrong
o Sandra is wrong -- if the heatsink is even *remotely* touchable
o Heatsink installed wrong - if the heatsink is cold (canted on black socket)

In Human Factors the guide was commonly 69oC as absolute limit of touchable
temperature before pain was such that it caused involuntary motor reflex. Contrast
with merely sensory pain like touching a 50oC laptop HD and noting "it hurts a bit".
Drug or Physical intervention could affect the latter (eg, TENS, Interferential) in a
limited way - such as 1oC at best on method-of-limits derived sensory threshold.

o User reports back - "it is extremely hot"
---- ok, but he *could* touch it -- you say 45oC, I say it could be up to 69oC
---- whatever, Sandra is miss-reporting 113oC
o User reports back - "it !@!?**! burnt me!"
---- ok, you might argue it could just be 46oC, I say it is around 69oC
---- whatever, the Heatsink is unsuitable for that product
o User reports back - "it feels quite cool"
---- ok, you might argue it could just be 45oC
---- whatever, I say the heatsink is canted on the S478 socket or CPU is not 113oC

The *threshold* of sensing pain at 45oC is good for hard-drives though.

You don't need to do any of this if you can eyeball the S478 socket:
o Even the Intel thermal solution will easily "cant" on the S478 side plastic rails
o Viewing the heatsink from the side will show whether it is canted on the receptacle

Personally I suspect a canted heatsink.
--
DB.

Dorothy Bradbury wrote:

Threshold of pain induced by heat is 45 degrees C; "69oC is the
touchable temperature limit before pain" is way too high.


Yes, that is the Psychophysics definition of the threshold of pain,
although figure varies according to usage, material & subject bias :-)
o Clinical Usage (eg, laser treatment):
---- Psychophysics list temperature threshold for pain as 45oC
o Design Material (eg, product liability):
---- Aluminium 45oC, Phenolic fan blade 61oC, Plastic socket 64-77oC
o Subject Bias (eg, conditioning or reward)
---- Temp perception of "pain" can be overriden beyond burning threshold
---- Coffee tasters report preference for hot coffee, despite it being 60oC
---- Yet 60oC is above the burn threshold, never mind above oral pain threshold

Yes. The issue here is the thermal capacity and thermal resistance of the
thing that's 'hot'. I.E. if it's of low thermal capacity, or has a high
intrinsic thermal resistance, then it (or it's surface) isn't 'that' hot
the instant it comes into contact with a cooler object.

Ok, back to the original poster:
o Poster reports "According to Sisoft Sandra, the CPU runs at 108-113'C"

Want a binary test for a) Sandra is wrong, or b) Heatsink installed wrong
o Sandra is wrong -- if the heatsink is even *remotely* touchable
o Heatsink installed wrong - if the heatsink is cold (canted on black socket)

Those are certainly good things to check but your logic tree is a bit
flawed. e.g. If the heatsink is canted so that it's cold then it is
certainly 'remotely' touchable but Sandra could likely be quite right.


In Human Factors the guide was commonly 69oC as absolute limit of touchable
temperature before pain was such that it caused involuntary motor reflex. Contrast
with merely sensory pain like touching a 50oC laptop HD and noting "it hurts a bit".
Drug or Physical intervention could affect the latter (eg, TENS, Interferential) in a
limited way - such as 1oC at best on method-of-limits derived sensory threshold.

o User reports back - "it is extremely hot"
---- ok, but he *could* touch it -- you say 45oC, I say it could be up to 69oC
---- whatever, Sandra is miss-reporting 113oC

Unless the TIM is poor, corrupted, cocked, etc., and Sandra is monitoring
the processor core temp diode.

o User reports back - "it !@!?**! burnt me!"
---- ok, you might argue it could just be 46oC, I say it is around 69oC
---- whatever, the Heatsink is unsuitable for that product

Or case airflow is insufficient.

o User reports back - "it feels quite cool"
---- ok, you might argue it could just be 45oC
---- whatever, I say the heatsink is canted on the S478 socket or CPU is not 113oC

The *threshold* of sensing pain at 45oC is good for hard-drives though.

You don't need to do any of this if you can eyeball the S478 socket:
o Even the Intel thermal solution will easily "cant" on the S478 side plastic rails
o Viewing the heatsink from the side will show whether it is canted on the receptacle

Personally I suspect a canted heatsink.

I would too except for him saying the suppliers claim it's a 'normal' temp
for their equipment.

On Thu, 29 Jul 2004 21:29:44 +0100, "Nel"
wrote:

I'm certain that my CPU (P4 3.06gHz) is running far too hot and is the cause
of my unstable PC. The manufacturer claims the temp is normal for this type
of system - which is more like a laptop than a desktop, as the motherboard
and cards are built in behind the TFT screen and the hardrive is built into
it's base.

According to Sisoft Sandra, the CPU runs at 108-113'C during normal
conditions an the top of the unit is hot to the touch. Can I rely on this
software's reading?

I've also heard that a P4 powers down when it reaches are certain temp to
prevent burn out - is there any software that can prove that this is
happening?

All help will be appreciated as this system has been in for repair 3 times
in 6 months and I'm thoroughly sick of it, and I want to look into the
possibility of getting a refund if they cannot supply me with a stable
system.

Thanks in advance


You have omitted (pretty much all of the important data).

What, exactly is this system? Hi-res pics (posted elsewhere,
not to the newgroup, then linked) or link to manufacturer's page
might be useful.

When system "has been in for repair 3 times", were symptoms same?
What did they tell you was wrong with it?

Does the system not have a software hardware monitor or bios
reading of temp and voltages? Forget Sandra, it makes a guess
that could be wildly off target.

Have you done the typical PC troubleshooting measures?
Have you checked all cables and connections, taken voltage
readings, examined heatsink-cpu interface, checked that fans were
working, ran memtest86 to check memory, or any/what other checks?

Those are certainly good things to check but your logic tree is a bit
flawed. e.g. If the heatsink is canted so that it's cold then it is
certainly 'remotely' touchable but Sandra could likely be quite right.

That's my point (further on).

I would too except for him saying the suppliers claim it's a 'normal' temp
for their equipment.

With thermal design limit well below 108-113oC, the supplier is making a
somewhat interesting comment - they may be just making a flippant fob off.

o Establish if the heatsink is seated correctly (and TIM material present/ok)
o Then email intel technical support once temperature is confirmed as 80oC
--
DB.

Dorothy Bradbury wrote:
Those are certainly good things to check but your logic tree is a bit
flawed. e.g. If the heatsink is canted so that it's cold then it is
certainly 'remotely' touchable but Sandra could likely be quite right.


That's my point (further on).

Yes, I know what you 'meant' but you still have, to the inexperienced, the
same condition leading to opposing conclusions

I would too except for him saying the suppliers claim it's a 'normal' temp
for their equipment.


With thermal design limit well below 108-113oC, the supplier is making a
somewhat interesting comment - they may be just making a flippant fob off.

Yes, I wondered about that too but, as I said to him, either they're wrong
about the temperature being 'normal' or they're wrong to dismiss the
problem, and I'm not sure which would be the worse offense.

o Establish if the heatsink is seated correctly (and TIM material present/ok)
o Then email intel technical support once temperature is confirmed as 80oC

Btw, you're presuming a desktop P4 and the temperature monitored being case
temp. Considering his description of the system being 'like' a 'laptop' in
form it made me wonder if perhaps it uses a mobile P4 and it's die temp
he's seeing (or, even if a desktop P4, die temp).

However, the temp would still not be 'normal'.

As kony pointed out, he's not provided enough information to know what he's
really got.