Arctic Freezer 7 Pro vs Blue Orb FX & Lapping. Me confused!

[Fruity]

System:
Q6600 OC to 3.4GHz
Vcore = 1.32V
Zalman GS1000 Case
Asus Rampage Formula Mobo

Temps before CPU lapping: Idle = 39°C Tcase; Full Load Prime95 Blend= 65°C Tcase


I was running my system using an Arctic Freezer 7 Pro CPU cooler, but last night I decided to lap my CPU as a push to anything greater than 3.4GHz is pushing my CPU over its 73°C design limit.

I've had a Thermatake Blue Orb FX sat on a shelf for several months, which was supposed to go in to the original system build, but for various reasons didn't. I figured that since I needed to remove the processor / mobo to do the lapping, I'd take the opportunity to fit the Blue ORb at the same time.

But....the temps after lapping and with the Blue Orb FX are about the same at idle (possibly a degree or two worse) and are several degrees higher under load. It is reaching 73°C Tcase within a few minutes of Prime95 running, at which point I'm stopping the test. The ambient that I was testing at was several degrees cooler than it is normally as well. Tjunc temps are at around 76°C - 81°C when Tcase is at 73°C. Under the old set-up the Tjunc tems would stabilise at around these temps.

Have I cocked something up or is the Blue Orb just not as capable as the Freezer 7 Pro? AS5 has been used with both coolers. If after the AS5 has 'settled' I get a 5°C improvement it will still mean my temps are still gonna be over what they were.
Also, does anyone know if it really matters if the Tcase goes above design spec providing the Tjunc temp don't get too hot?????

I'm pretty confident that the IHS is now properly flat as I checked it before, during and after lapping with an engineers straight edge - it is about as flat as I have ever seen Shiny as well



I'd do a back to back comparison by putting the Arctic Freezer back on again, but that unfortunately means taking the mobo out again which I don't really want to do unless I have to!!

Should also say that the Blue Orb does a fantastic job of cooling the NB chipset, bring its temp down by 5°C under load! Not that it was ever in danger of overheating....

 

[RJR]

Nice case, Nice Mobo, Nice processor.

Thermal specs:
Q6600 B3: 62.2c
Q6600 G0: 71.0c

Don't really understand the choice of cooler without a window case. It seems to me that the only reason you would buy a cooler like that would be to show it off through a window.

I would suggest a better cooler (True,1283,CCF,etc.) to handle a higher OC, or just stay at 3.4 Ghz which isn't bad at all.

 

[Fruity]

Thanks for the reply RJR

Point taken on the choice of cooler. I originally chose it because its specs indicated that it should have done a better job of cooling than the Arctic 7 - at which point I don't care how many pretty lights its got
At the time I couldn't find any reviews comparing it directly to the Arctic 7 (or indeed anything else) but it fell within budget so I got it.
Not good reasoning i know!!

At this point I'll put the Arctic 7 back in which will hopefully bring my temps down again.

I would stay at 3.4GHz, but I can't. I KNOW that the system has more to give and dammit I'm gonna draw every last MHz out of it So far I'm only using an 8x multiplier, i reckon with new 1066MHz RAM I can get it over 3.8GHz. A bit of a pointless pass-time I admit, but it keeps me off the streets lol Thanks for the pointers for a new cooler. I'll probably replace it when I get the new RAM sometime in the summer. Might do some more research this time as well!

Just as a side note, I'm not getting on too well with the Zalman case - it's almost impossible to keep the hard drives cool. The PSU is mounted at the bottom of the case, and the point at which you should be able to bring cool air up from underneath is packed full of wires from the PSU. There's no way to force air in from the front either. Not a good design for an air OC'd system. Am really looking for a good alternative case that has better air flow availability and hot-swap drive bays. Ideas??

 

[B-Unit]

Did you also lap the HSF? If you really want to see gains from your lap job, you need to flatten both. When I did my Opty165+AF64, I got 2-3C out of it, not as much as I had hoped but not bad overall.

EDIT: Just saw your second post, and I know the feeling. Had always hoped for 3.0Ghz out of the one I lapped, the lapping finally let me get to 2.9Ghz.

 

[Fruity]

I didn't lap the HSF, but it was already fairly smooth and flat. Because of the size, weight and shape of the HSF it would have been a tricky task at best. I might attempt to do the Arctic 7 when it goes back in (and that's no quick job!). Will post a couple of photos when i do the change just to show off (or leave myself open to ridicule!)


Thanks for you post.

Fruity

 

[Fruity]

Put some pics up of the lapped HSF

http://www.tomshardware.co.uk/forum/page-251522_11_0.html

Enjoy!

 

[jsc]

The ACF7P is pretty flat on the base. And 3.4 GHz. with a Q6600 is very good with what is basicly an economy cooler. I stopped at 3.3 GHz. with mine. I did get to 3.6 GHz. with a TRUE. I could get my Q6600 to post at 3.8 GHz. but it wasn't stable - stability defined as 24 hours, Prime95, small fft.

In you quest for more speed, I think that you will find that the memory is not your limiting factor.

 

[Fruity]

Oh ok. that's interesting. What do you see as the limiting factor?
The NB has plenty of voltage and temp remaining. The CPU has lots of volts left (but needs better cooling).
I figured i couldn't get a stable overclock when i attempted 3.6GHz because the DDR2-800 RAM I'm currently using was maxing out. Replacing it with 1066 ( i thought!) would give me more head-room

Any thoughts?

Saying that, I did run 3.8GHz for long enough to run 3DM06, which gave an impressive 3000+ extra points compared to 3.4GHz. Enough to make a difference to my normal PC usage of data crunching. Temps went over recommended during that though.

 

[plenge]

hello.. i just bought a Blue orb FX.. and its 36-37 degrees hot.. is that normal?

 

[Fruity]

What are your:

load temps?
processor?
processor voltage?
is it overclocked?

 

[plenge]

my processor is a quad core Q6600 2.4GHz

dont really know about the voltage and load temps, and its not overclocked either

 

[plenge]

Oh. i found the voltage. Its: 1.176V

 

[Fruity]

For an idle temp at stock speed it does sound a little warm although load temps are the important bt.
Download and install Realtemp and monitor temps while running Prime95.

Don't allow the case temp (Tcase) to exceed 70°C or the core temps (Tjunc) to exceed 90°C for any core. That does give a little slack before their actual max design limits.

http://www.techpowerup.com/downloads/1325/Real_Temp_3.00.html for realtemp and

http://files.extremeoverclocking.com/file.php?f=103 for Prime95

 

[plenge]

hmmm.. do you think it'll work with a little more Cooler Paste? i personally think i put too less cooler paste on the new cooler when i installed it. (seemed to be alot more on the standard cooler i got with the processor) allthough i know too much isnt good for the pc either, since its gonna warm up the processor even more.

i only have one other fan in the pc, thats on the side of the case. but i bet that the blue orb should be able to keep it way more cool alone than it is now.

 

[ouch1]

Plenge,

Too much thermal paste can be worse than not enough. Too much can cuase it to act as an insulator and cause your cpu to retain heat. Where as too little will just cause inefficient conduction. I have seen people use too much and cause thier cpu's to run 5-10C higher than they should run.

-ouch1

 

[Conumdrum]

This is the crux of the matter. Too much or too little is bad. If temps aren't right, then you need to redo it. Once you learn after trial and error whats right it usually takes one try.

The paste is a 'helper' between the minute differences (in 1000's of an inch) in smoothness of the CPU and HS. To put in in perspective, a sheet of newspaper is a bit less than 3/1000 of an inch. If ther eis a gap of just 1/2 of 1/1000 of an inch, the heat will not transfer.

The paste isn't as efficient as a pure metal to metal perfect contact. It needs to be as thin as possible.

To check you gently remove the cooler and check the TPaste pattern. It should be thin and cover the complete top of the CPU and bottom of the HS. Too little and it won't cover the full area. Too much and it blocks heat.

It's a learned experiance. You learn fast when you have to remove the mobo, undo the screws each time or make sure the pins are fully seated.

I learned the hard way. You MUST inspect the bottom of the mobo if you use pins, they are notorious for not seating right, and the only way to check is look at the bottom of the Mobo.

I use a watercooling block now, it has a base thats attached to the bottom of the mobo, I don't have to inspect the bottom anymore, just make sure the screws were tightened evenly in a cross pattern. Heck, that alone takes over 5 minutes, if not 10.

Random ramblings.