Congratulations on your new Q6600 box. I haven't read all the thread
but I thought I'd contribute my ideas (though they may be redundant
wrt. others' comments I haven't seen, or things you've discounted already).
1) There's a specific way to apply AS5 for specific use on the Kentsfield
quad-cores like the Q6600; the diagram / instructions for that process
has a unique link on AS's web site. It is not the same instruction as to
application for a dual core or other kind of CPU. The direction of the
bead of paste has to match the direction the cores are laid out under the
IHS side by side. Double check that you've used a process that's
going to give at least as good paste coverage results as that process.
2) After having used AS5 on the first of my Q6600 boxes
(with a lapped ThermalRight Ultra Extreme 120 HS, NOT-lapped CPU IHS),
I found AS5 to be a bit thick / viscous and I'm not particularly confident
that it'd spread as well as optimum even given their special technique
and using excess amounts of paste. For my next build I think I'll use
a very liberal amount of AS Ceramique instead since I think it'll spread
more widely over the quad core CPU cores and IHS and require less
"break in" time to achieve stable results.
3) I'm running Q6600G0 on a P5K-Deluxe around 35-38C IDLE at stock
volts/frequency, and around 42-48C load temperature also at stock
volts/frequency, and I believe my AS5 isn't applied well for the above
reasons as well as that I had difficulty mounting the HS which resulted in
some smearing as I twisted it around, and in fact the TRUEX120 is loose
so I know it should be tighter for best results.
4) Some temperature monitoring utilities have a known 15C read error on some
Core2 series CPUs; this is an error due to the fact they changed the reference
"Tjunction" temperature that the RELATIVE temperature sensor readings are
RELATIVE TO in some of the chips; some programs written to use 100C
and others use 85C and the "correct" value depends on your particular
chip model; I forget the details of which is which; check the Intel forum on
xtremesystems or google around for tjunction + conroe + 85c or something
like that and you'll find the details.
5) Update to the latest stable BIOS that has explicit CPU support
for your CPU stepping and model, and double check the BIOS's temperature
at BIOS IDLE matches closely your OS utility programs' temperature reading,
and also verify that the version of your temperature utility matches the
right Tjunction value for your stepping / CPU as above. It could be a BIOS
bug or utility bug is just reading all your temperatures as 15C too high maybe??
Clearly a Q6600 should run hot, but I would not expect 75C at either IDLE or at
load anywhere under 3GHz.
Debug the temperature stuff at stock multi/frequency/volts just to avoid
complicating your analysis.
6) Some BIOS versions of some motherboards (I've no idea about yours)
overvolt/overclock BY DEFAULT; this isn't at all what you want for initial
testing / debugging. Ensure it's either at STOCK volts/multi/frequency
(around 1.16V to 1.3V core full load; x9 CPU multi; 266 MHz master clock);
if the BIOS has a bug or overclock setting these could be much too high
and may be causing your problems.
7) When / if you remove the HS carefully look for the actual imprint
of the HS and IHS in the paste; see how far it spread, and how uniformly;
there shouldn't be any unusually thick (as if untouched) or unusually
thin islands due to very nonuniform contact. You can use a flat clear
glass from a small picture frame to print from your IHS or HS base
with a single drop of water in the middle. If the drop spreads out
to cover the whole IHS/HS on the glass then you know it's pretty flat.
Also do the light seeping through the razor edge gap thing behind
a bright light.
8) Check to make sure your CPU socket's loading plate (the hinged one)
isn't bent up or something that might make bad HS contact.
9) check to make sure your "motherboard" temperatures are
reasonable relative to your room temperature to help ensure
the BIOS / utility is reading reasonable temperatures, though
that COULD WELL be INDEPENDENTLY ok/bad to your CPU
reading.
10) Try manually forcing the CPU / clock / volt settings;
try 1.25Vcore, 266 MHz master clock, x9 multi, and see if
the temps cool down.
If you know your nominal VID/Vcore values from your CPU utility
then of course use that Vcore (slightly higher or slightly lower)
if available, otherwise, 1.15 to 1.25V should be "about right";
using 1.15V shouldn't hurt if it's too low for your particular CPU,
though it could be unstable if it needs a little more (but I doubt it).
11) If you want to really find out what's up, then after you check
the temperature reactions to the above trials, now UNDERCLOCK the
CPU by changing the CPU multiplier to x6 instead of x9
(also with manual low Vcore and 266 MHz) ; if THAT doesn't radically
drop your temperature readings, either your HS is just not working/mounted
right, or your temperature utility is really out to lunch, or your BIOS / MB
is really not setting the expected parameters for the CPU since there's
NO WAY it'd run hot at x6/266Mhz/1.15Vcore.
Good luck!
Originally posted by: ArchAngel777
Ok, decided to post a few pictures of my setup to see if someone can give me any last pointers.
My
case
My
Tunig Tower
I'll see if I can explain this post a little more thoroughly than my previous ones.
If I run my CPU @ stock speeds - Q6600 @ 2.4Ghz 1.3v (actual is 1.24) - I idle around 54c and load around 75c. This is with speed step and voltage step enabled.
If I run my CPU @ overclocked speeds - Q6600 @ 3.2Ghz 1.38v (actual is 1.34) - I idle around 58c and load around 78c.
The issue here doesn't seem to be related to any overclocking as this CPU runs very hot without doing anything to it. What is worse, though, is that PWM (NB) progressively gets hotter and hotter and it doesn't matter if I am at stock speed or overclocked it just runs damn hot. The longer I run Prime Maximum Heat, the hotter it continues to get. PWM hit 85c earlier today with an ambiant temp around 27c.
So, here are my plans, feel free to take a look at the pictures and tell me if anything is off. Keep in mind that the Tuniq Tower fan is moving air towards the rear exhaust fan and I also have an exhaust fan above it. To the right of my graphics card you will notice an intake fan that blows on the graphics card, slightly around the NB (Blue Heatsink) and then gets exhausted out the rear and top. My Hard Drive is on the lower chamber and any heat is exhasted through the PSU, which is also at the bottom.
Here is what I feel the issue could be
1) Defective Tuniq Tower. This is a real possibility because I have read a few people on the internet that had a bum Tuniq Tower. I might try and use the Intel Heatsink + Fan and see if temps are the same or lower.
2) Sensor is wrong - though I doubt this is the case... The temps raise and I can feel the heat coming out the exhaust fans.
3) Concave CPU and or Heatsink - I am not thrilled with the idea of lapping my CPU. I know it is safe, but it voids the warranty. I eye balled both the bottom of my Tuniq AND my IHS and didn't notice any issues with it. But, I know that isn't a scientific method... I really should have gotten a razor blade to test it with. In any case, lapping my CPU is to be a last resort...
4) It is possible I just have a bad CPU. Maybe the IHS is not bonded to the four cores correctly or something. No way of really knowing for sure.
I guess I have no real way of proving #2 and #4, so I might try #1 and if that fails I might actually RMA this CPU. Though I am not sure NewEgg will appreciate me doing that because it is running hot. I have done everything in good faith to get these temps down and I am pretty much out of ideas.
Note: People keep telling me that I must have the Tuniq Tower mounted wrong - please, don't keep saying that. I have remounted it twice and I am very confident in my ability to apply thermal grease (AS5) and mount a HeatSink. Tuniq Tower is the one of the easiest heatsinks to mount, at least with LGA775. Turn each one full turn them move to then next, and next, and the next, etc... until all of them are tight and bingo, that is all there is. Nothing difficult about that. I can pickup the motherboard with the Tuniq Tower and it does not move. So, again, the Tuniq Tower is mounted properly.
Note: I am also not the only one with abnormally high temps for a Q6600. Been doing a lot of searching, hours actually, and find that there are several reports that others have been running into this. They also had reseated and reseated and reseated their CPU coolers over and over to no avail. Some users reporting 85c under load! I am pretty certain mine would climb up to those temps if I really let it go for 6 hours or so...