• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Good Cooling!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

eobard

Give me a break Senior
Joined
Jul 12, 2001
Just thought I'd post this for all to read. It started by my looking on a web site at a bunch of altered cases. Cases altered to make the systems within cooler. Many cases had a grill covering a fan on the side of the case relative to the position of the CPU. It occured to me that that kind of set up would bring in cooler air to the whole case but not necessicarily directly to the CPU, a lot of re-circulating of heated air within the case would still get to the CPU. I decided to try an experiment. I removed the door from the "component side" of the case, I traced its size onto a sheet of cardboard and then cut it to size. Next I figured where the CPU would be in relation to that sheet of cardboard. I cut out a circle there and the duct taped a tube there. The tube was a can of orange juice with the metal top and bottom removed. I sized the tube so that when the cardboard sheet was put in place of the metal door the tube would be right over the CPU fan / heatsink combo. Once I had it lined up properly I tested my computer with both the metal door and with the cardboard door with the "blow hole". I expected the temperatures to be much lower with the tube because at least 90%+ of the air the heatsink's fan put onto the heatsink would come from outside the case at room temperature. With all other factors being as identical as possible here are the results I got:

Conditions.........................| Blowhole....| No Hole
-----------------------------------|------------------|-------------------
Idle 10 minutes................. | 37c/98f.......| 39c/102f
-----------------------------------|------------------|-------------------
Need4Speed 4.................| 39c/102f......| 43c/109f
10 minutes running
-----------------------------------|-------------------|------------------
SiSoft Sandra burn in.......| 40c/104f.......| 45c/113f
15 minutes running


5 degrees celcius / 9 degrees ferenheit under heavy load. No water blocks, no pelitiers, not even an extra fan!

Hope it helps someone.
 
OP
E

eobard

Give me a break Senior
Joined
Jul 12, 2001
Updated info: I had a friend of mine who's a machinist cut a precision hole in the case door and I hot glued a plastic speaker tube in it. (You know how some people have big wooden boxes in their car trunks with lots of speakers, covered in black speaker carpet? Take a close look at one sometime, most will have, aside from speakers, holes cut in them with only a plastic tube fitted into the hole. Thats the kind of tube I'm refering to.) The tube is centered on the fan and the tube goes deep enough to leave only about 3-4mm clearence between the tube and the fan. This is far more efficient than the cardboard temporary case door and tube that I was using. I didn't use the exact same method of comparison but my previous tests weren't scientific either. Here are the results:

Boot temp: 30c! (I didn't post the initial boot temp before but in both cases then it was 35c)

Idle temp: 33c (37c with cardboard tube, 39c with standard door before)

NFS5 8 minute running temp: 37c Last time I tested with NFS4 running 12 minutes but I just downloaded the demo of Need for Speed 5 and wanted to try it. Also I figured that the newer game would be more demanding and result in higher temps, it didn't. Last time I got 39c with cardboard tube and 43c with standard door.

SiSoft Sandra Burn in test: Last time I ran it for 15 minutes and got 40c with cardboard and 45c with standard case door. This time I ran it for FOURTY SIX MINUTES and only got the temp up to 39c!!!!!!!

Now I'm thinking about sealing the gap between the tube and outer edge of the fan, perhaps I'll get the temp down even more.

Six degrees celsius cooler under harsher testing conditions!! Cost to me: Three dollars (Canadian) for the speaker tube, ten cents for the amount of hot glue used and ten minutes of a machinist's time (free for me, aren't friends wonderful?) Lets see a peltier be that efficient!! I recomend this to all. One warning though: direct access to the CPU fan is quite the cat hair magnet! Keep that fan clean.
 

outhouse

Senior Member
Joined
Mar 11, 2001
Location
Auburn California
Ive been running a duct for a long time and its the best mod that has helped the most out of any that ive done i was able to overclock with high CV with a cheap HSF and had great results now with a good HSF [with a Delta] i'm even getting better results I wonder why more people dont use this mod with the cost being just rite foe most overclockers :)

goodluck all