My e6400@3.3Ghz is running too hot!

[gamephile]

Hi everyone,

I've been wrestling with heat problems ever since I built this latest machine. As you can see from my sig I'm pushing the CPU, but not anymore than the Ultra 120 should be able to handle. My case is an Antec P150 with a 120mm exhaust fan sitting right in line with the Ultra 120.

While I used to have temps like those listed in my sig, after even a minute or two of TAT load, my temperatures measured by TAT shoot to around 68 degrees. My Tcase temps aren't much better, around 63. Usually there's a pretty big discrepancy between the two.
I idle around 42-45. My environment has changed and is a little warmer than where I had it previously, but not 10-15C warmer! Although I have head a rumor/rule of thumb that for every degree increase in ambient you gain two in cpu temp.

I've lapped both the cpu surface and (to a light extent) the surface of the Ultra 120. I used AS5 as my thermal grease. I keep doubting myself as to whether I did a good job lapping, or applied the AS5 correctly. I'm going to consider buying some shin-etsu x23 soon, but I have doubts of that being my saving grace.

Another worry is that since my case is upright, the Ultra 120 seems to hang down a little bit. Could this be affecting it's contact with the CPU? Obviously it's mounted pretty tight, but if this were to bend the motherboard in a weird direction maybe it would affect temps.

Ugh, I'm becoming very frustrated by an otherwise fantastic build. Any help you could offer would be appreciated.

 

[Nnyan]

Have you taken off the Ultra 120 and made sure you had good coverage with the AS5? I have the U120E and I have to tell you that it does like having decent air flow. I shut off the intake fans on my Lian-Li PC-7S and there is a noticeable difference in temps. In fact I ended up reversing the exhaust fan so that I have two 120mm intakes and one 120mm exhaust at top (cut out a bigger blow hole then the little one there). That really made a significant difference in my case temps.

Having said all that my E6420 idles in the low to mid 30's and hit's 59-60 under heavy stress testing. I think I read that the older chips run a bit hotter but I don't think it should be that big a diff. What do you mean it "hangs down a bit"? I think maybe you don't have that mounted properly.

 

[gamephile]

Yeah I'm considering adding another fan, either in the front or doing a dual mount on the Ultra120 itself.

As for the drooping I noticed, stick your arm out straight with your palm parallel to the ground. Imagine your hand is the Ultra120 and your wrist joint is where it meets the CPU. Ever so slightly dip your hand down towards the ground. This is what I mean.

Another note, I'm currently trying to lower my Vcore but this damn chip is unforgiving. I lowered it in the bios to one step below 1.4 and Orthos fails after 33 minutes. Sigh.

 

[Accord99]

I wouldn't exactly call 68C with TAT workload too hot. TAT is by far the hottest running application publicly available, it runs much hotter than something like dual Prime95 which itself runs a C2D to higher temperatures than most typical applications. Plus, depending on the stepping you have, the danger point only starts at either 85C or 100C as measured by TAT.

 

[gamephile]

Yeah I logically I know this, but I still feel like I'm doing something wrong. I have a chip that the Ultra120 should be able to take in stride and when I see other users with lower temperatures I feel like there's more I could do.

 

[zagood]

I'd definitely add at least one intake to the front of your P150, I've got two on mine (1 Akasa Amber @ 12v, 1 Panaflo M1A on a fan controller). I removed the filter, it was too restrictive for the Akasa.

-z

 

[gamephile]

I should've mentioned earlier that I have a 100mm intake fan blowing across my 2 hard drives. They probably heat the air enough that the increased airflow doesn't help the CPU cooler.

 

[n7]

Okay seriously, i am ridiculously sick of C2D & temps posts, sorry...

People have been asking the same freakin' questions since the CPUs came out last August...

C2Ds run HOT (since they measure temps on the die)!

This is not abnormal at all!

Ambient temps play a HUGE factor in CPU temps.

Another thing, different batches of C2Ds run VASTLY different temps.
My F batch E6600 ran cooler with way more vcore than my current B batch with much less vcore!


Too hot is when you're running over 80C (measured w/ iTAT or CoreTemp)
Then it's time to get worried!

At this exact moment, my E6600 is idling @ 50C using the best air cooler out there (though not with the best fan for temps in this case).
Orthos will hit around 70C after running for a bit, & this is with a mere 1.27V under load, since my CPU happens to be one of the hotter ones, & my case isn't great for airflow.

You're running 1.4V & getting only 70C load lol?

That's nothing...if anything, i'd be pushing it farther.

Sorry for this little rant, but let me assure you, 70C is nothing for C2Ds.

As i mentioned, 80C is where you need to start getting worried.

 

[gamephile]

Thanks for the reply N7. I know that every other post on these boards is about C2D temps, I guess I got selfish. I'll pass on your experiences whenever they're necessary

 

[n7]

Sorry about the rant as well...i just see the same concerns brought up all the time.

The problem i see also in every thread is other posters stating temps are bad, because theirs run cooler, or etc.

Except unless you have the exact same case, exact same ambient temps, & exact same CPU, there's no way to generically compare cooling results unfortuntely. (Since all of those play a large factor)

In my own situation, i actually found around a 10C difference at idle & load between my F batch E6600, & my B batch E6600 with the same vcore.
Now most people assume that's because of poor mounting, etc. etc, but it wasn't.
On the same note, the F batch could do around 3.6 GHz max (this was with my Coolit Freezone TEC water cooler), whereas the hotter B batch E6600 could do around 3.9 GHz max.

IOW, to sum it up, YMMV!

One thing I would suggest is seeing what temps are like with the case side open...it'll give you a decent idea of if the case airflow is poor

 

[God Mode]

thanks N7..refreshing to see that. People constantly complaining about their cpu temps for personal home computing are displaying borderline to severe OCD complex. Do these same people wipe the dust off their fan daily and reapply the TIM whenever their cpu seems to be a 'C or two higher?

 

[Nnyan]

I would be a bit concerned about the Ultra 120 drooping, it almost seems that one side is not tightened down enough. I had a concern about that with my ultra 120 extreme and they basically told me to do "hand tight" which to them means use a screwdriver with little/no pressure until it will not go anymore (unless you bore down on it which you shouldn't do).

One of the reasons I got the E6420 was b/c of the reports that it runs cooler then the plain XX00's. I've got mine rock solid at 3.2Ghz and I think I'm really close to tweaking it to 3.4Ghz (so far it will run about 1.5 to 2 hours of Orthos before it error's out. But I have not finished tweaking it yet). Anyways it idles at mid-30's and I have not been able to get it to go above 60.

 

[gamephile]

All of my screws are tightened until the screw stops, so no worries there. The droop is very slight, I think I'm making a mountain out of a molehill, it's like the ultra120 just kinda stretches it's limbs a little. EDIT, actually I think this probably happens to everyone, I just notice it because with my combination of a DS3 and the P150, there is a metal support bar that's part of the case that runs right above the right edge of the ultra120. When the ultra120 droops it just touches and springs off the metal bar.

And yeah the actual silicon makes a temp difference. I got one of the allendale based C2D's and I think these run a little hotter and are a little less forgiving about necessary voltage than their conroe based brothers. At this point I'm going to ride it out until a Quad core chip is is dirt cheap and decently overclockable. I'm glad you got yours to 3.4, but I keep telling myself that the effort required to push my machine an extra 90Mhz just isn't worth it. Regardless, I still want to make it as cool running as I can.

 

[SimMike2]

I guess I have the opposite question, in that my 4300 core temp doesn't run hot at all.

This is with using the stock heatsink (at least until my other cooler arrives.) I can barely get it to go over 40c. Now granted the case is open and it isn't that warm in Seattle today. But still, I am running at 3GHz with my Gigabyte 965 DS3. To get it to run stable, I upped the voltage to a setting I found on the Internet, which CPU-Z shows as 1.44. I think the setting in the BIOS was 1.48125. I'm using speedfan and picking Core 0 to monitor. I think the temp is being read correctly, as it goes up and down dramatically depending on load.

I haven't really put it to the full test, running orthos or prime 95, until I get the better heatsink. But so far I'm impressed. Right now I'm running the longest Super PI test, which should take about twenty minutes.

At least according to SuperPI, this things smokes my Opteron dual core running at 2.5GHz. Whereas the AMD was lucky to get 34 seconds with the 1M test, this 4300 cranks it out in 19 seconds. That's almost twice as fast!

The long SuperPI test just finished. It was just under 19 minutes. My Opteron dual at 2.5 took 34 minutes for the same test. That is a wipeout. I hope these result in fast video encode and stuff like that as I plan on doing quite a bit of video editing in the next month or so.

 

[nyker96]

Originally posted by: SimMike2
I guess I have the opposite question, in that my 4300 core temp doesn't run hot at all.

This is with using the stock heatsink (at least until my other cooler arrives.) I can barely get it to go over 40c. Now granted the case is open and it isn't that warm in Seattle today. But still, I am running at 3GHz with my Gigabyte 965 DS3. To get it to run stable, I upped the voltage to a setting I found on the Internet, which CPU-Z shows as 1.44. I think the setting in the BIOS was 1.48125. I'm using speedfan and picking Core 0 to monitor. I think the temp is being read correctly, as it goes up and down dramatically depending on load.

I haven't really put it to the full test, running orthos or prime 95, until I get the better heatsink. But so far I'm impressed. Right now I'm running the longest Super PI test, which should take about twenty minutes.

At least according to SuperPI, this things smokes my Opteron dual core running at 2.5GHz. Whereas the AMD was lucky to get 34 seconds with the 1M test, this 4300 cranks it out in 19 seconds. That's almost twice as fast!

The long SuperPI test just finished. It was just under 19 minutes. My Opteron dual at 2.5 took 34 minutes for the same test. That is a wipeout. I hope these result in fast video encode and stuff like that as I plan on doing quite a bit of video editing in the next month or so.

Sim, 1.48 is gonna fry your chip at full load. Be careful, not saying you can't do it, but I think running with stock HSF and 1.48 isn't very good combo, if you don't need that much volt tune it down a little.

 

[jakegub]

I was having really bad performance with my Thermalright SI-128 and all I did was remove it and add a lot more thermal paste. My heatsink/cpu must be very badly lapped because I could see it was only making contact in a very few places. When I put on more than I thought I needed, I am now at 1.55V and never higher than 71 at load. I say re-do your thermal paste. Also, 70 at load at 1.44V is WAY too high for the circumstances. You aren't going to hurt the processor, but at 1.44V I think I Loaded at about 58.

 

[VanH]

Sorry to hijack the thread but this one seemed the closest to my recently completed project. I have to say this was a joy to build and clock; assembly took 6hours, os, drivers and prelim overclocking the last couple of days. E6420 clocked at 3.4 rock solid stable, over an hour of orthos no problem, superpi 32M at 15min. Going to switch out the cpu fan; Think I can move it to 3.6+; but it's very sweet right here, low noise system.

http://img255.imageshack.us/im...034ghzoverclockyo4.jpg

 

[Nnyan]

OK, I played around with my CPU and I've found that with the Ultra 120 extreme that if you don't put in enough TIM it can make a huge difference in your temps. For example I covered the bottom of the heatsink and the CPU with a full coat of AS5. It was a medium thin coat and based on my previous experience what I thought to be more then enough.

Carefully put everything back together and ran Orthos. After about an hour I was at 69-72. I was very surprised, I didn't think it would make that much of a difference. So I re-sat the CPU two more times just to make sure (and checked out the TIM coverage) and same results.

So I put back the additional TIM (an extra thin line) and my temps peaked at 60-62 (which is 2 degrees higher then I was getting before so I may re-do the TIM). I would suggest you try adding some additional TIM to see how that goes.