Help me get to the bottom of this temp issue (c2d)

[ui]ICEMAN

[ui]ICEMAN

[ui]ICEMAN

Soldato
Joined
18 Oct 2002
Posts
2,612
Location
Florida/UK
Hi,

Been having some issues with a 6600 I have and the Evga 680i SLI board. Using TAT and Core Temp I seem to be getting very high temps with a very very small increase in voltage. At 1.325v 3.4ghz under Orthos I'm getting ~58-60c on each core in TAT which I'm fine with. However, when stepping the voltage up even slightly to get 3.6ghz (1.4v and anything more) I end up with core temps rocketing to over 90c.

I've checked heatsink contact and I've even lapped the IHS to no avail. I'm using a Tuniq tower and I have a Zalman 9500 which I've tested too but still no joy. I'm tempted to whack it in the Vapochill and be done with it but I get the feeling there's something deeper amiss here. Perhaps poor contact between the cores and the IHS? I'm no newbie to overclocking, I've been playing with cascades, LN2 and phase change for many years and even on air, I've never seen temperatures rise so quickly with such a minor change in vcore.

Anyone have any ideas?
 
Hi

It could be the actual voltage being used is much higher than you've set in the bios, a few mainboards do this :(

thats the only explanation i can think of

regards.
 
Check delta between idle and load. If its a BIG jump REALLY quickly then could be poor IHS contact with the core.
Do both cores show the same jump?
If the systems stable @ 3.6/1.4v then ignore the temps :)
 
System is no longer stable at 1.4v 3.6 so I assume the temps are probably correct. Temp rise is incredibly rapid when load is applied, within 5-6s they're up from the idle temps of ~35-40 to 80+ @ 1.4v.
 
Last edited:
Concorde Rules said:
Hmm, Bottom on TT120 warm/hot?

If not I'd say core--> IHS contact is ****.

:(
Click to expand...

Yup, or the software's lying through its ass.

1.4v is quite low for 3.6ghz stable, maybe stick another cooling method on it and see if that improves temps.
 
Heatsinks are all relatively cool, warm under load and definitely hot with the fans off. Its looking like poor IHS - core contact isn't it :(

Its only a stop gap chip til my Kentsfield arrives anyway but still, I'd be tempted to keep this one if I could sort this issue. Warranty is void anyway, may as well remove the IHS and whack it under the vapo.
 
[ui]ICEMAN said:
Heatsinks are all relatively cool, warm under load and definitely hot with the fans off. Its looking like poor IHS - core contact isn't it :(

Its only a stop gap chip til my Kentsfield arrives anyway but still, I'd be tempted to keep this one if I could sort this issue. Warranty is void anyway, may as well remove the IHS and whack it under the vapo.
Click to expand...

From what you've said it sounds like there's reasonable/decent contact with the IHS, since the sinks getting warm. Likewise the sink getting hot with no fan suggests there's a reasonable amount of heat getting through to the sink. What are temps like comparing fan on vs fan off?
You'd expect to see differences in temps between the 2 cores if contact was poor, due to the die area's, it's pretty unlikely the "poorness" of contact is uniform across both cores.
 
What's interesting is that I've heard reports of several others, both from this forum and others who have received 6600's this last week and had the same issues with temperature. In fact there's one literally right below this thread. Bad batch perhaps? Temps with no fan go up by a few c, nothing dramatic but I have reasonable airflow around my case.
 
Was very poor til I lapped the IHS, literally only one side of the CPU would leave an imprint with a fine layer, its nigh on perfect since the lapping though. I didn't have anything above 600grit to hand last night though so it was only a rough lapping to get the surface even.
 
its very maybe possible that the HSF's just can't cope quickly enough with the power change.

It realy sounds like they are becoming the insulator(heatpipes).

they usualy have a rating of maximum thermal capacity - i do wounder if your close to saturation.

off the top of my head, those heatpipes are usualy around 26W each
 
Others are running substantially more than 1.4v though on similar or the same cooling, without the insane temperatures, I'm really puzzled by this one.
 
best guesstimate is around 120W, the HSF should be able to cope with 200W so can't be that.
if we quickly do the maths and leave alrge margin for error:

Total output 120W
Thermal resistance 0.2°c/w

120x 0.2 = 24°C

I would even consider a 25° swing reasonable.

Working from your figures even if we consider ambient to be 40°c, your CPU has to be close to 200W output, or your thermal efficency is total parp , but on 2 coolers - doubtful.




The Only thing I can think of is your mainboard, can you check your voltages when under load to see if you get a larg drop off ?

Maybe someone with similar mainboard and a 6600 closer to you could help you out.
 
Im also struggling with temps on my 6600 under water. 30idle and 40 load at 1.5v but 40 idle and 60 load at 1.6v which seems a big difference under water. I lapped cpu and hasnt made much difference. Only thing i can think is the ihs contact is bad but its soldered on so contact is generally very good
 
ted34 said:
Im also struggling with temps on my 6600 under water. 30idle and 40 load at 1.5v but 40 idle and 60 load at 1.6v which seems a big difference under water. I lapped cpu and hasnt made much difference. Only thing i can think is the ihs contact is bad but its soldered on so contact is generally very good
Click to expand...

1.6v is quite a lot of voltage, remember this is over 2 cores..
Seeing similar (ish) temps at 3.9ghz and 1.6v on water here tbh, so they dont seem that far off.
The CPU's do kick out quite a bit of heat once you up the voltages and clockspeeds, similar to the AMD X2 series, once you bung above say 1.45v into the chips temps start to rocket.
I'd be less worried about your temps than the poster of this thread about his.
 
An Interesting read is here about the DTS - I wounder if some of the callibrations might be way out.(http://www.overclockers.com/articles1378/)

Also a worthy download is TAT ([size=-1]Intel Thermal analysis Tool).
You can grab it at[/size][size=-1] http://www.techpowerup.com/downloads/392/mirrors.php, its good as you can set each cores load independantly and watch the variation of the temps.

the only other temp program that reads the DTS correctly for sure is Coretemp. http://www.xtremesystems.org/forums/showthread.php?t=103638

Since they are digital, mainboard or voltages have NO IMPACT on the readings.
[/size]
 
Me..same issue:confused:

When I built this system at stcok speed and volts memps were quite low for air cooling...mid 20's idle and about 40 full load..that was using Gigabyte's easytune5.
I overclcoked and have got a stable system at 3.5Ghz with 1.4V on the CPU in BIOS. Easytune reads a CPU voltage of 1.38. Im using a Scythe Ninja rev B.
My temps idle using Intel TAT are mid 40's and when full loading on both cores using Orthos hits about 70C. Ive tried reseating the cooler many times and have even hacked a bit of the mounting bracket cos I though it was resting against some motherboard capacitors (using a DS4)....still no change in temps.
I bought my CPU from here just under 2 weeks ago from here...do you think we have got a bad batch that dont seem to control the temps very well...albeit very stable at 3.5Ghz. :confused: :confused:

Also why does the motherboard monitor software read a lot less on the temps?? up to 10degrees less?

edit...when fully loading my heatsink never gets very hot...only slightly warm to touch.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom