Woot -- 4.01Ghz Air Cooled! (Q6600)

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

JAG87

Diamond Member
Jan 3, 2006
3,921
3
76
Originally posted by: Laminator
Originally posted by: JAG87
Originally posted by: njdevilsfan87
Originally posted by: aigomorla
Cooling requirement needed for this speed and settings to be stable.
http://i125.photobucket.com/al...aigomorla/IMG_0721.jpg

:D

Now that's just not fair.


try this then

1
2


and its still not possible to have 4Ghz Prime95 stable. maybe you can run 3d mark, or fold, but you will not get it prime stable. and even if you do, its probably in such a form factor that is not suitable for 24/7 use. lets be realistic, unless you have phase change cooling where the amount of voltage needed for stability drastically decreases, 4Ghz is just a number.

Hey, JAG87, you have Corsair XMS2 C3's? I know you are probably running them 1:1 at 800MHz C3, but have you tried any overclocking past 800MHz?

yes I run 1600 FSB and 1:1 with the memory at stock right now, but I have tried overclocking it.

this ram will do just about anything as long as you keep the stock timings of 3-4-3-9 and 2T command rate. I have ran 830 mhz on stock volts (2.2v), 850 mhz at 2.3v, 900 mhz at 2.4v, and I stopped there. However if you switch to 1T you need about 2.4v to keep it stable even at 800 mhz. It can easily run 3-3-3-8 with 2T at 800 mhz, but there is no gain in performance for doing that. 1T brings the greatest benefit and given the voltage needed, its not a good idea for 24/7.

if you dont care for CAS3, this memory can do anything at CAS4, and still gets a slight bump from CAS5. at 4-4-4 it can run 1067 mhz with 2.4v, and I believe it can probably hit 1100 at 5-5-5 with 2.4v.

I wouldn't dare run more than 2.4v through any DDR2 memory for day to day usage. Thats why I stop there. The IC under the hood is D9GKX, and its probably one of the best DDR2 ICs ever produced together with D9GMH. Yes there are kits that kill this memory as far as speed overclocking (look at the OCZ Flex XLC 9600C5 and the Dominator 10000C5), but this memory was binned for low latency, and thats exactly what it does. 900 mhz at CAS3 is no joke, without the need for active cooling! You better have good airflow though, or youll end up with 2 overcooked bacon strips.


Originally posted by: Nathelion
Now this is what you'd really need to keep a quad at 4 GHz
http://www.spikedhumor.com/art...eral_Oil_Computer.html


Thats honestly a good idea. The motherboard would benefit so much from that. The NB would be a lot cooler, and the hardest thing to cool which is the PWM, would be a lot cooler too. but whos going to clean my hardware when I want to sell it...

 

Nathelion

Senior member
Jan 30, 2006
697
1
0
Mineral oil ends up eventually frying the system though, just to warn you. The mineral oil itself is not conductive, but it will get contaminated with various conductive particles over time. Also, it doesn't dissipate heat very well - if you run your system for a few hours, the mineral oil will get hotter and hotter until it's VERY hot and the computer crashes. I suppose you could pump the oil out into some kind of "cooling grid" or run it in a split tube with water on the other side (and just switch the water out when it's too hot) or some other complex solution, but somehow, I doubt it's actually worth it:)
Besides, it'd be a hassle to upgrade, lol. And yeah, the resale value might not be that great. Still, it would be fun to take an older computer that you don't care about and do this to it and then overclock it like crazy:)
 

JAG87

Diamond Member
Jan 3, 2006
3,921
3
76
Originally posted by: Nathelion
Mineral oil ends up eventually frying the system though, just to warn you. The mineral oil itself is not conductive, but it will get contaminated with various conductive particles over time. Also, it doesn't dissipate heat very well - if you run your system for a few hours, the mineral oil will get hotter and hotter until it's VERY hot and the computer crashes. I suppose you could pump the oil out into some kind of "cooling grid" or run it in a split tube with water on the other side (and just switch the water out when it's too hot) or some other complex solution, but somehow, I doubt it's actually worth it:)
Besides, it'd be a hassle to upgrade, lol. And yeah, the resale value might not be that great. Still, it would be fun to take an older computer that you don't care about and do this to it and then overclock it like crazy:)

good points. thats probably why its not a common cooling practice...
 

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org
Originally posted by: Nathelion
Mineral oil ends up eventually frying the system though, just to warn you. The mineral oil itself is not conductive, but it will get contaminated with various conductive particles over time. Also, it doesn't dissipate heat very well - if you run your system for a few hours, the mineral oil will get hotter and hotter until it's VERY hot and the computer crashes. I suppose you could pump the oil out into some kind of "cooling grid" or run it in a split tube with water on the other side (and just switch the water out when it's too hot) or some other complex solution, but somehow, I doubt it's actually worth it:)
Besides, it'd be a hassle to upgrade, lol. And yeah, the resale value might not be that great. Still, it would be fun to take an older computer that you don't care about and do this to it and then overclock it like crazy:)

You can use water pumps and filters to keep it clear.

The whole point of most immersion projects though is to eliminate condensation when cooling below ambient. However, mineral oil gels too early. Flourinert is too expensive. If this guy had good results, you can bet your ass he wouldve posted them in that video.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,587
10,226
126
Originally posted by: aigomorla
im talking about prime95 mark. On small FFT's.
Bonic she stays at 1.51V straight though without dropping. For some wierd reason when i prime at small FFT's the vdroop gets larger and larger by the hour.
Interesting. Perhaps the mosfets are overheating??