Stock Core 2 Duo Question

[jjw]

I just finished my Core 2 Duo (E6600) build. I am running everything stock at this point.

I'm sure this is a dumb question, here goes...

Is my CPU running warmer than 'normal'? At Idle in Win2K the temp is 55 C, Fan 1025 RPM. I ran the 'stress test' included with the motherboard, the temp went up ~1C, the fan speed went up to ~1500 RPM.

I have seen temps posted much lower than this, I'm wondering if I screwed up the heatsink mounting or if there is some thermal regulation mechanism going on.

Specs:
E6600 (w/Stock heatsink - mounted w/arctic silver5)
Intel D975XBX M-board (flashed to latest bios)
2 GB 677MHz Patriot Memory
GeForce 7900GS
Some leftovers for the rest.

 

[knupxfon]

seems kinda high to me. My x6800 runs at 32c idle and 40c load. What hsf and thermal paste are you using?

 

[jjw]

Stock HSF and Arctic Silver 5

I disabled the Fan AutoSpeed in the bios, Temps down to 53C, Still high, Fan at 1780 RPM

 

[Mex]

I think you need to try reseating your heatsink.
The Intel stock cooler is notorious for failing to seat properly on the CPU.

Do a search before making a new thread, since many people have been having temperature issues.
http://forumz.tomshardware.com/hardware/built-Conroe-System-crazy-temps-ftopict199845.html
http://forumz.tomshardware.com/hardware/Core-Duo-E6600-configuration-ftopict199705.html

 

[jjw]

With the Fan speed disabled in the bios the temp dropped to about 53C and the fan speed is stuck at 1780 RPM. The system has only been on for several hours so far...

My stock E6300 is about 49-52C idle, and about 58C full load. Wait a little for AS5 to kick in properly.

I am a little worried that the stock Heatsink is too rough for the AS5 to work well. The machining marks were pretty well defined after I removed the stock goop from the cooler.

--edit
FAN is at 2250 RPM now.
CoreV is 1.294

 

[jjw]

Jack,

Ambient is likely around 75 F (maybe less).

As a correction my Fan is pegged at 2250 RPM and temps are 52C -CPU, 38C MB location 1, 36C MB location 2.

I'm not really worried, just don't want to 'futz' with something that is OK as is. Just want to make sure I didn't do something wrong, seeing that Intel's Heat sink is prone to being put on wrong!

Hopefully the AS will 'kick' and the temps will come down.

 

[levicki]

I am a little worried that the stock Heatsink is too rough for the AS5 to work well. The machining marks were pretty well defined after I removed the stock goop from the cooler.

Well the stock HSF is pretty flat. IHS on the CPU on the other hand is sometimes concave so I always advise people to stick with stock thermal pad if they intend to use stock HSF but nobody seems to be listening.

Everytime I read a complaint about high temperature it starts with:

I removed thermal pad, put AS5 and installed stock HSF...

Face it people, you cannot improve stock HSF by changing default TIM (thermal interface material). You can only screw it up. There were several articles which compared Intel thermal pad and various aftermarket thermal paste efficiency and each and every one of them said pad is better and that you should not remove it unless you won't use stock HSF, do some research next time.

 

[cableuser]

I am a little worried that the stock Heatsink is too rough for the AS5 to work well. The machining marks were pretty well defined after I removed the stock goop from the cooler.

Well the stock HSF is pretty flat. IHS on the CPU on the other hand is sometimes concave so I always advise people to stick with stock thermal pad if they intend to use stock HSF but nobody seems to be listening.

Everytime I read a complaint about high temperature it starts with:

I removed thermal pad, put AS5 and installed stock HSF...

Face it people, you cannot improve stock HSF by changing default TIM (thermal interface material). You can only screw it up. There were several articles which compared Intel thermal pad and various aftermarket thermal paste efficiency and each and every one of them said pad is better and that you should not remove it unless you won't use stock HSF, do some research next time.

Thank you for this info. I am getting my core 2 duo parts in tomorrow and I would have done the same thing others are doing. That is take of the pad and put on as5.

Back on topic:

You said the machining is a bit ruff so maby you could sand paper that sucker down. Start out with a coarser grit and than move to a more finer grit. Than maby you could apply some AS5 and have it work better than it is now. I think the process is called lapping? I may be wrong though. I have tried it on my 478 heatsink and it worked. I got about a 2-3c drop in temps. Of course my heatsink was machined fairly well to begin with so your case may be different.

 

[Artmic]

Actually your cpu at idle runs hot, what do you use to measure the temps?
Unless your room temp is like 30C ? if it is like 23C to 25C it is way too hot imo. It should be around 10C lower at least, that is a weird temp.

Make sure lots of air is blowing into your case. i switched my 2 sidepanel fans for Panaflo M1's running at full blast and temps went down by 9C just because of that.

It don't matter how good or bad a Stock intel heatsink is, if it is blowing 40C+ hot air on top of a cpu/motherboard nothing will help i guess.

 

[levicki]

Thank you for this info. I am getting my core 2 duo parts in tomorrow and I would have done the same thing others are doing. That is take of the pad and put on as5.

You are welcome. Anyway, try it with the pad -- you can always remove it and put some AS5 but first check what you have with stock parts. If you mount it correctly and get poor result then you start worrying.

You said the machining is a bit ruff so maby you could sand paper that sucker down.

That is not important. What is important is whether the surface is flat or not. It is easy to test it. Spread thin layer of simple white thermal paste over the HSF surface and press it evenly against a glass plate. If you get evenly covered surface then you don't have to worry about it. The tiny scratches are filled with paste anyway.

However, you might want to perform the same check with the CPU before installing it and if you do not get an evenly covered square on the glass surface then you will hopefully understand why the stock pad is so thick.

 

[clue69less]

I am a little worried that the stock Heatsink is too rough for the AS5 to work well. The machining marks were pretty well defined after I removed the stock goop from the cooler.

Well the stock HSF is pretty flat. IHS on the CPU on the other hand is sometimes concave so I always advise people to stick with stock thermal pad if they intend to use stock HSF but nobody seems to be listening.

Everytime I read a complaint about high temperature it starts with:

I removed thermal pad, put AS5 and installed stock HSF...

Face it people, you cannot improve stock HSF by changing default TIM (thermal interface material). You can only screw it up. There were several articles which compared Intel thermal pad and various aftermarket thermal paste efficiency and each and every one of them said pad is better and that you should not remove it unless you won't use stock HSF, do some research next time.

Thank you for this info. I am getting my core 2 duo parts in tomorrow and I would have done the same thing others are doing. That is take of the pad and put on as5.

Back on topic:

You said the machining is a bit ruff so maby you could sand paper that sucker down. Start out with a coarser grit and than move to a more finer grit. Than maby you could apply some AS5 and have it work better than it is now. I think the process is called lapping? I may be wrong though. I have tried it on my 478 heatsink and it worked. I got about a 2-3c drop in temps. Of course my heatsink was machined fairly well to begin with so your case may be different.

If levicki's claim that the IHS is sometimes concave is correct (I have no reason to doubt it and have heard the same claim from others) then lapping it could help. Hand lapping tends to result in a convex surface and I've long thought that fact could be as important in the temp reductions people see after lapping as much as is improved smoothness.

 

[jjw]

I am a little worried that the stock Heatsink is too rough for the AS5 to work well. The machining marks were pretty well defined after I removed the stock goop from the cooler.

Well the stock HSF is pretty flat. IHS on the CPU on the other hand is sometimes concave so I always advise people to stick with stock thermal pad if they intend to use stock HSF but nobody seems to be listening.

Everytime I read a complaint about high temperature it starts with:

I removed thermal pad, put AS5 and installed stock HSF...

Face it people, you cannot improve stock HSF by changing default TIM (thermal interface material). You can only screw it up. There were several articles which compared Intel thermal pad and various aftermarket thermal paste efficiency and each and every one of them said pad is better and that you should not remove it unless you won't use stock HSF, do some research next time.

Your sig states your system, could you quote your core temps, and if you have the stock HSF or not and possibly any other details. There has only been one quote of temp, similar to mine with a x6300.

Respectfully, I have not seen the articles you have described. All the articles I have seen start basically with we removed the stock TIM and applied AS for a 'fair' comparison of the heatsinks. In addition, (being a Arctic Silver drone...) any previously applied TIM will affect future TIM applications. I've seen nothing to back this up though.

Does the stock heat sink just suck and these are 'normal' temps or are they 'much better' if the stock TIM is used.

prozac26:
Arctic Silvers site quotes a 2-5C temp drop over the breakin period.

I am reseating my heatsink right now. When I pulled the heatsink the AS was covering about 75% of the heat spreader. (about the same as the pics on Artic Silvers site; not as even and circular as their photo though) Then I'll give it the full 200 hr breakin.

 

[Guest]

Nobody never mention this, but when I put up Pc with LGA775(wich is quite often), I always put the CPu and the HSF first. I do it by putting the anti-static bag in my left hand and then the motherboard on top of the anti-static. I then push the 4 pins in place. Afterwards I flip the mobo to make sure the four little black thingy got all the way thru. This is a simple yet effective way to make sure your heatsink is seated properly. I found that mounting the motherboard ans then the HSF doesnt always give proper results.

Since you seam to have some time I suggest popping the motherboard out of the case and make sure all the black thingy are all the way thru!

 

[jjw]

I just replaced my Stock cooler with a AC Freezer 7 Pro. Temps dropped instantly to 45C, not idle, but under stress

I decided to use the stock TIM on the cooler. I will note that there was a significantly higher amount of force retaining this new heatsink over stock.

My guess is that there is not enough pressure exerted by the stock cooler to squeeze AS to a thin layer. The AS with the stock cooler did not appear to squeeze thin. (Using the method recommended by AS to apply a small amount at center then installing the HS without pre-spreading the AS)

 

[jjw]

Yes, I cleaned both HS and CPU before applying AS. Ambient is ~25C (77F).

I've disabled the fan in bios, it only makes a small difference (2-3C). It seems that the fan runs slow, only up to 1800 RPM, shouldn't it be more like 2500 rpm.

The new cooler dropped the temps by almost 10C! I'll give it a couple days and then I'll ponder lapping the new cooler. Like I said in my previous post, I think the stock cooler does not exert enough pressure, and in addition is likely not very flat/smooth.

 

[jjw]

Thanks for all the help.

The moral of the story, the stock cooling solution is lousy, really lousy. I have yet to see anyone say they have cool temps with the stock solution.

 

[aequalsb]

I'm Running the "lousy" HSF. I'm also enjoying very low CPU temps! I did however replace the OEM paste and used AS. Maybe that's why...

Edit! Holy Shit.. I didn't even look at those numbers.... LOL I'll mess w/ the wiring and get back to ya...

 

[jjw]

Nice try, Unless your PC is sitting in a freezer, there is no way in hell that you have core temps that low. You'd have to be breaking the second law of thermodynamics if you're doing it on the stock cooler. Try this on for size...

It's prettly likely that you're not being entirely honest, or you have some bad thermal data, your core voltage looks a little high for stock BTW. You really should check your PS as it looks like you're getting about .9 volts out of the 12 v line (it is amazing the system even boots)

 

[derek2006]

Your sensors arent reporting that right. There is no way that your cpu is at 1 Celsius.

 

[aequalsb]

I must have some faulty stuff. I looked in the bios and the cpu is reporting 43-45C. I tried speedfan and that reports really crazy data. Like 7000 rpm fan speed... What's that about. Should i take this MOBO back cause of a faulty sensor?

 

[aequalsb]

Alright. I used the Gigabyte MB thingy and that one seems to have good data. see below.

How do those numbers work for the ole laws of thermodynamics?

 

[jjw]

That looks better. Stock cooler huh, I guess there must be some wide variation in how well they work.

I've seen numerous reports of bad sensor data...

 

[levicki]

About thermal pad and stock HSF versus aftermarket stuff this is what I was referring to:
http://www.tomshardware.com/2004/11/03/5_cool_p4_coolers_protect_your_p4_from_heatstroke/page3.html

What is interesting is the result for thermal resistance. The cooler reaches just 0.37 K/W with the heat-conducting paste supplied with it, which is applied to the cooler in a similar way to a heat-conducting pad. When conventional heat-conducting paste is used, the thermal resistance rises to 0.51 K/W! For this reason, we recommend that you use the accompanying heat-conducting paste.

More here:
http://www.tomshardware.com/2004/11/03/5_cool_p4_coolers_protect_your_p4_from_heatstroke/page10.html

The boxed cooler from Intel achieves the best score when the heat-conducting paste supplied with the cooler is used. With conventional heat-conducting paste, Intel's cooler slips into the middle of the pack.

For Core 2 Duo use Core Temp to measure temperature.

Putting mainboard on a solid surface and pressing is not wise thing to do because you can short some pins on the lower side if you accidentaly bend them by pressure. If you already have to do it out of the case it is better to position HSF in place and then using both hands press the notches two by two diagonally with your thumbs while at the same time you hold the back side of the mainboard with your fingers. Just be sure to be properly grounded beforehand.

Anyway, stock HSF is not bad. It works great but people are usually afraid to push hard enough for the four notches to click in place. Sometimes one of them turns a bit while pushing and it won't latch.

Another thing is concave IHS I mentioned before. Lapping it won't help much except to lose the warranty. If you are so unlucky to have one, get aftermarket cooler and that should fix it.

 

[jjw]

Levicki,

Thanks for the articles! I may even have read them but my memory has not served me well

As for my High temps. I think I have it narrowed down to this:

- My Core2Duo has a concave heatspreader
- 'new' Aftermarked Cooler has a Concave spot on center (seen after I lapped the cooler.)
- ArcticSilver5 works poorly at large gaps
- My ambient / case temps are high (77F Ambient/90 case)
- Stock cooler fan spins slow, <1800 RPM even with MB control shut off
- I'm running Win2K, that doesn't have SpeedStep drivers by default, Idle and Loaded are very close (~1-2C)

This is what I have done to come up with this conclusion:

1) Stock cooler, Artic Silver 59C [70% coverage, Irregular pattern of AS after HS removal]
2) Stock cooler, Artic Silver 55-57C [75% coverage, circular pattern of AS after HS removal]
3) AC Freezer 7 Pro, Stock TIM 45C [100% coverage of IHS]
4) AC Freezer 7 Pro, Artic Silver 49C [75% coverage, circular pattern]
5) Lapped to 600 Grit AC Freezer, Artic Silver 44C [-haven't removed HS yet]

When I can find some higher grit sandpaper, I'll lap the AC Freezer more. But for now it is flatter and slightly smoother than stock.

I am still curious I why the stock TIM works better than AS. The AS 'should' have a higher conductivity than the other stuff. I wonder if a heavier application (i.e. 100% coverage) would work better than the 75% that 'AS corp' recommends when the heatsinks are not lapped and polished?

:idea: Technical Analysis:
The P4 cooler from the article is quoted as having a thermal resistance of .37W/C or .51W/C. If the C2D is assumed to be ~65W then the deltaT from core to air would be 24-33C or 9C cooler with the stock TIM! As for the AC Freezer Pro (.17W/C -grain of salt MFG spec) deltaT would be 11C. This should have lead to a drop in temp of 13-22C. If my case temp was 30C, then 55C for temps was to be expected.

 

[levicki]

Thanks for the articles! I may even have read them but my memory has not served me well

You are welcome. Seems that too many people has forgotten that article. Another usefull source of information for cooling and quiet systems is www.silentpcreview.com/. Check their articles about coolers and PSUs.

My Core2Duo has a concave heatspreader

Quite possible.

'new' Aftermarked Cooler has a Concave spot on center (seen after I lapped the cooler.)

You probably picked low quality brand.

ArcticSilver5 works poorly at large gaps

True.

My ambient / case temps are high (77F Ambient/90 case)

You should try to solve that. No cooler can help if you have 30°C in your case.

Stock cooler fan spins slow, <1800 RPM even with MB control shut off

That is not slow. Default is 900 RPM. 1800 RPM is way too loud for my taste.

I'm running Win2K, that doesn't have SpeedStep drivers by default, Idle and Loaded are very close (~1-2C)

Then you should install Windows XP because 2K doesn't use new powerstates.

When I can find some higher grit sandpaper, I'll lap the AC Freezer more. But for now it is flatter and slightly smoother than stock.

You just make sure that you don't make it concave or convex. Lap it on a thick large piece of glass.

I wonder if a heavier application (i.e. 100% coverage) would work better than the 75% that 'AS corp' recommends when the heatsinks are not lapped and polished?

You can try to mount HSF without any TIM and measure if there is any difference and how much. That would be interesting thing to find out.