Intel Core 2 Temperature Issues

[Gerry_uk]

Hi,

I'm experiencing problems with Intel Core 2 E6800 temperatures on my
ASUS mainboard, but part of the problem is I don't even know if the
values I'm seeing reported on screen are correct or not!

I've created a quick web page with a screen shot and details here

<http://www.xp20.dircon.co.uk/hardware/>

Does these readings make sense?

 

[Fred]

Hi,

I'm experiencing problems with Intel Core 2 E6800 temperatures on my
ASUS mainboard, but part of the problem is I don't even know if the
values I'm seeing reported on screen are correct or not!

I've created a quick web page with a screen shot and details here

<http://www.xp20.dircon.co.uk/hardware/>

Does these readings make sense?

Well from what I understand the core 2 duo cpu's have an old style sensor
diode under the integrated heat spreader (The BIOS uses that for temp
measurement) and 2 digital sensors one above each core.
It's not uncommon for the measurements taken from the diode and core sensors
to vary.
Intel also provide a utility to access the core sensors and apply some
stress testing.
Available for download here
http://anonforums.com/builds/TAT.zip
HTH

 

[Phil]

Looking at those temps, I'm inclined to think that the Probe II reading is
not to be trusted. As for the CoteTemp readings, they're a bit high if those
readings were taken at idle. I would re-check that the heatsink is properly
mated to the motherboard, and making good contact with the CPU heatspreader.
In any case, a good aftermarket cooler (with artic silver 5) is a much
better cooling solution than the stock HSF.

-phil

 

[Gerry_uk]

Hi Phil,

Looking at those temps, I'm inclined to think that the Probe II reading is
not to be trusted.

This is the strange thing; the recent BIOS update I applied to this
board states that it's purpose is to provide more accurate temparature
readings. The old BIOS gave lower readings.

As for the CoteTemp readings, they're a bit high if those
readings were taken at idle.

Yes it was Idle for two hours.

I would re-check that the heatsink is properly
mated to the motherboard, and making good contact with the CPU heatspreader.
In any case, a good aftermarket cooler (with artic silver 5) is a much
better cooling solution than the stock HSF.

Yup, this is exactly what I tried to do and failed

Here's the ASUS readings as I went along:

Intel cooler, old ASUS BIOS [CPU 50 idle, 60 load]
Intel cooler, new ASUS BIOS [CPU 55 idle, 65 load]
Zalman cooler, new ASUS BIOS [CPU 61 idle, 62 load]

I originally thought the problem was either the Intel cooler is no good
or the thermal contact between CPU and heatsink was flawed. I took off
the Intel cooler and the thermal compound was all squashed down BUT it
had some small holes in it, as if there wasn't enough.

I assumed this was the problem and got the Zalman cooler and the Actic
Silver 5. I took out the whole MB, reseated the CPU, cleaned everything
and let dry. Put quite a lot of Actic 5 on the CPU, bolted down the
Zalman, then used a torch to look at the connection. Everything looks
flat and some compound has squashed out the sides.

I put everything back assuming it was all fixed now, and total
disappointment as I see the idle temp is WORSE than before!

I note the Zalman heatsink is COLD. Could it be the junction is still
flawed?

What I did next was disconnect the CPU fan and power up again. This time
the heatsink became extremely hot, extremely quickly, so to me it seems
the heat transfer _is_ occurring?

I'm out of ideas

 

[Gerry_uk]

Intel also provide a utility to access the core sensors and apply some
stress testing.
Available for download here
http://anonforums.com/builds/TAT.zip
HTH

Thanks Fred, I'd like to try this, but "anonforums" doesn't sound like
Intel! Do Intel have a link to it?

 

[Barry Watzman]

They make sense but they seem high. My E6600 (running completely stock,
not overclocked) is reporting temps in the mid-30's (C), and I believe
it, the heatsink isn't even warm. I'd wonder if the heatsink is mounted
properly. This is the best CPU I've ever used.

 

[greysky]

It does seem to be a bit on the high side for running 'stock'. I have the
E6400 and a Zalman cooler as well, and running unstressed the temps are
around 41C - 42C, using Core Temp. Overclocked, the cores go to as high as
71C before I chicken out and lower everything. One thing that may mess up
your heat transfer is applying too much Arctic Silver 5 - I noticed you
saying you used a lot of it. You really want only a thin layer spread
evenly over the heatspreader. God luck in figuring this problem out.

Hi Phil,

Looking at those temps, I'm inclined to think that the Probe II reading
is not to be trusted.

This is the strange thing; the recent BIOS update I applied to this board
states that it's purpose is to provide more accurate temparature readings.
The old BIOS gave lower readings.

As for the CoteTemp readings, they're a bit high if those readings were
taken at idle.

Yes it was Idle for two hours.

I would re-check that the heatsink is properly mated to the motherboard,
and making good contact with the CPU heatspreader. In any case, a good
aftermarket cooler (with artic silver 5) is a much better cooling
solution than the stock HSF.

Yup, this is exactly what I tried to do and failed

Here's the ASUS readings as I went along:

Intel cooler, old ASUS BIOS [CPU 50 idle, 60 load]
Intel cooler, new ASUS BIOS [CPU 55 idle, 65 load]
Zalman cooler, new ASUS BIOS [CPU 61 idle, 62 load]

I originally thought the problem was either the Intel cooler is no good or
the thermal contact between CPU and heatsink was flawed. I took off the
Intel cooler and the thermal compound was all squashed down BUT it had
some small holes in it, as if there wasn't enough.

I assumed this was the problem and got the Zalman cooler and the Actic
Silver 5. I took out the whole MB, reseated the CPU, cleaned everything
and let dry. Put quite a lot of Actic 5 on the CPU, bolted down the
Zalman, then used a torch to look at the connection. Everything looks flat
and some compound has squashed out the sides.

I put everything back assuming it was all fixed now, and total
disappointment as I see the idle temp is WORSE than before!

I note the Zalman heatsink is COLD. Could it be the junction is still
flawed?

What I did next was disconnect the CPU fan and power up again. This time
the heatsink became extremely hot, extremely quickly, so to me it seems
the heat transfer _is_ occurring?

I'm out of ideas

 

[John Lewis]

Hi Phil,

Looking at those temps, I'm inclined to think that the Probe II reading is
not to be trusted.

This is the strange thing; the recent BIOS update I applied to this
board states that it's purpose is to provide more accurate temparature
readings. The old BIOS gave lower readings.

As for the CoteTemp readings, they're a bit high if those
readings were taken at idle.

Yes it was Idle for two hours.

I would re-check that the heatsink is properly
mated to the motherboard, and making good contact with the CPU heatspreader.
In any case, a good aftermarket cooler (with artic silver 5) is a much
better cooling solution than the stock HSF.

Yup, this is exactly what I tried to do and failed

Here's the ASUS readings as I went along:

Intel cooler, old ASUS BIOS [CPU 50 idle, 60 load]
Intel cooler, new ASUS BIOS [CPU 55 idle, 65 load]
Zalman cooler, new ASUS BIOS [CPU 61 idle, 62 load]

I originally thought the problem was either the Intel cooler is no good
or the thermal contact between CPU and heatsink was flawed. I took off
the Intel cooler and the thermal compound was all squashed down BUT it
had some small holes in it, as if there wasn't enough.

I assumed this was the problem and got the Zalman cooler and the Actic
Silver 5. I took out the whole MB, reseated the CPU, cleaned everything
and let dry. Put quite a lot of Actic 5 on the CPU, bolted down the
Zalman, then used a torch to look at the connection. Everything looks
flat and some compound has squashed out the sides.

Mistake !!

You must apply a very thin layer of Arctic 5. Enough to fill any
un-eveness and no more! Seeing it squish out the sides is not a good
sign. For a CPU and heatsink both with polished-smooth surfaces and no
obvious visible indentations, 3 or 4 very small beads smoothed flat
with a finger such that the CPU header is visibly entirely covered
with a THIN gray layer of Arctic 5 should be quite sufficient. The aim
is to get the CPU header and heatsink core in <<as close a contact as
possible>> with all un-eveness voids filled and no more than that.
Compared with the heat-transfer efficiency of the heatsink slug and
the CPU header, Arctic 5 is a much poorer heat conductor.

I put everything back assuming it was all fixed now, and total
disappointment as I see the idle temp is WORSE than before!

I note the Zalman heatsink is COLD. Could it be the junction is still
flawed?

No. You possibly used far too much Arctic 5, or you have not properly
clamped down the Zalman and have insufficient pressure, or both.

What I did next was disconnect the CPU fan and power up again. This time
the heatsink became extremely hot, extremely quickly, so to me it seems
the heat transfer _is_ occurring?

Sure, but the question is: How efficiently ?

John Lewis

 

[Phil]

First, too much AS5 is bad, for two reasons. First, too much will actually
inhibit efficient heat transfer from the CPU to the HSF. Second, AS5 is
conductive, so it cause problems if any seeps out onto the motherboard's
circuitry. I would remove the HSF, clean it and the CPU, and then apply a
very small amount of AS5 - about the size of a small grain of unccoker rice.
Spread it evenly across the surface of the heatspreader. I'm still inclined
to think that the whatever sensor the BIOS and Probe II are reading is
faulty. There's no way your CPU is idling anywhere near 60C. In any case,
it's the core temps that are the most important, so go by whatever CoreTemp
and TAT tell you. They should both give almost identical results, as they
both read thge digital thermal sensors on the CPU die. I would consider
RMA'ing the board, especially if you can get a replacement cross-shipped to
you.

Did your core temps improve at all, after switching to the Zalman cooler?

-phil

Hi Phil,

Looking at those temps, I'm inclined to think that the Probe II reading
is not to be trusted.

This is the strange thing; the recent BIOS update I applied to this board
states that it's purpose is to provide more accurate temparature readings.
The old BIOS gave lower readings.

As for the CoteTemp readings, they're a bit high if those readings were
taken at idle.

Yes it was Idle for two hours.

I would re-check that the heatsink is properly mated to the motherboard,
and making good contact with the CPU heatspreader. In any case, a good
aftermarket cooler (with artic silver 5) is a much better cooling
solution than the stock HSF.

Yup, this is exactly what I tried to do and failed

Here's the ASUS readings as I went along:

Intel cooler, old ASUS BIOS [CPU 50 idle, 60 load]
Intel cooler, new ASUS BIOS [CPU 55 idle, 65 load]
Zalman cooler, new ASUS BIOS [CPU 61 idle, 62 load]

I originally thought the problem was either the Intel cooler is no good or
the thermal contact between CPU and heatsink was flawed. I took off the
Intel cooler and the thermal compound was all squashed down BUT it had
some small holes in it, as if there wasn't enough.

I assumed this was the problem and got the Zalman cooler and the Actic
Silver 5. I took out the whole MB, reseated the CPU, cleaned everything
and let dry. Put quite a lot of Actic 5 on the CPU, bolted down the
Zalman, then used a torch to look at the connection. Everything looks flat
and some compound has squashed out the sides.

I put everything back assuming it was all fixed now, and total
disappointment as I see the idle temp is WORSE than before!

I note the Zalman heatsink is COLD. Could it be the junction is still
flawed?

What I did next was disconnect the CPU fan and power up again. This time
the heatsink became extremely hot, extremely quickly, so to me it seems
the heat transfer _is_ occurring?

I'm out of ideas

 

[Phil]

TAT is very difficult to find. It's not even available from Intel's site.
The link provided is what you want. Just beware that it was made for earlier
pentium CPUs, and has not been updated to recognize the conroe line of CPUs.
It works fine, but detects conroes as Pentium Ms.

-phil

 

[Gerry_uk]

Mistake !!

You must apply a very thin layer of Arctic 5.

OK thanks; everyone is saying the same, I'll do a total clean up and try
again.

 

[Gerry_uk]

TAT is very difficult to find. It's not even available from Intel's site.
The link provided is what you want. Just beware that it was made for earlier
pentium CPUs, and has not been updated to recognize the conroe line of CPUs.
It works fine, but detects conroes as Pentium Ms.

OK, TAT all set up now, it's a nice program! It agrees exactly with the
Core Temp program. ASUS Probe seems to be way off. Note that ASUS Probe
agrees with the reading in the BIOS (which is expected)

 

[Barry Watzman]

He's right .... in fact, that stuff is electrically conductive, and if
it gets on things or goes where it's not supposed to, it can destroy a
system.

The correct amount of thermal compound is very small ... about the size
of one to three grains of (uncooked) rice.

 

[Gerry_uk]

Hi,

OK thanks; everyone is saying the same, I'll do a total clean up and try
again.

I cleaned everything (it took a while) and then applied a very thin
layer of AS5 compound to the CPU, put everything back. Unfortunately,
there's no perceptible difference

I notice this Zalman CNPS8000 cooler says it's for Pentium D and Pentium
4, but surely Core2 is the same?

I also notice the Zalman instructions for the Intel CPU say to use
"generous" amount of Zalman thermal compound whereas the instructions
for AMD say to use a "thin layer" - that seems strange!

I also notice two of the LGA pins on the mainboard socket look odd, I
didn't notice this before?

The top of my CPU is only just above the load plate, but it is above - I
don't think the load plate is getting in the way.

 

[Larc]

| First, too much AS5 is bad, for two reasons. First, too much will actually
| inhibit efficient heat transfer from the CPU to the HSF. Second, AS5 is
| conductive, so it cause problems if any seeps out onto the motherboard's
| circuitry. I would remove the HSF, clean it and the CPU, and then apply a
| very small amount of AS5 - about the size of a small grain of uncooked rice.
| Spread it evenly across the surface of the heatspreader.

Everything is fine until the last step. I have found putting the AS5 on the
center of the CPU (but NOT spreading it any), then seating the heatsink and
allowing that contact to do the spreading works best.

Something else that helps but hasn't been mentioned is preparation of the CPU
and heatsink surfaces. Clean both surfaces well with isopropyl (not rubbing)
alcohol or acetone. Let dry. Then apply some AS5 to the heatsink surface, put
your finger inside something like a plastic sandwich bag, and use your finger to
thoroughly rub the AS5 all over the heatsink surface area (you want to protect
your finger AND keep oil from your finger away from the heatsink). Then clean
the AS5 off the heatsink surface with a lint-free paper towel or a piece of
crushed newspaper. Be careful not to touch the heatsink surface with your
finger(s). The purpose of all that is to get AS5 into tiny holes and crevices
on the heatsink surface that may not even be visible to the naked eye. That
way, there shouldn't be any dead air spaces in your CPU/heatsink connection, not
even very tiny ones.

I've heard this said many times: with AS5, less is more!

Larc



§§§ - Change planet to earth to reply by email - §§§

 

[Phil]

What exactly do you mean by "odd"? If any of the pins are bent or missing,
RMA that sucker. Regarding the CPU placement in the socket, it should be
visibly obvious whether or not the CPU is properly seated.

As for your temps, if nothing has changed, that seems to confirm that the
sensor read by the BIOS and Probe II is faulty. All things considered, I
would RMA the board.

-phil

 

[Black Shuck]

Hi,

I'm experiencing problems with Intel Core 2 E6800 temperatures on my
ASUS mainboard, but part of the problem is I don't even know if the
values I'm seeing reported on screen are correct or not!

I've created a quick web page with a screen shot and details here

<http://www.xp20.dircon.co.uk/hardware/>

Does these readings make sense?

My results on a P5B, with E6600. I also see a noticable difference
between the PCProbe and the Core Temps.

http://picasaweb.google.com/mark.gillespie/Junk

 

[kony]

I also notice the Zalman instructions for the Intel CPU say to use
"generous" amount of Zalman thermal compound whereas the instructions
for AMD say to use a "thin layer" - that seems strange!

It seems more wrong than strange. Don't put faith in
aftermarket product instructions, there is never a case
where there should be more than enough compound applied to
just barely make a *complete* interface. If the heatsink
base were rough, it will require far more than if smooth.

Generally on a very smooth heatsink the amount of grease
needed is less than a grain-of-rice sized blob, applied only
to the middle (not like spreading a cake where one tries to
completely cover the entire top surface).

 

[Gerry_uk]

Hi,

My results on a P5B, with E6600. I also see a noticable difference
between the PCProbe and the Core Temps.

http://picasaweb.google.com/mark.gillespie/Junk

Your offset looks similar to mine, but obviously your temps are much
lower and your fan speed is lower too; the only difference is you are
using E6600 and I'm using E6800.

 

[Andrei Outkine]

I am experiencing 50-55C on my E6400 in idle/low load mode with ASUS
P5B, Intel stock fan and Antec Sonata II with an air duct.

What is considered to be "normal" for these processors? And what's the
highest safe temperature?