fan airflow direction

[Dumbledore]

Which is better: have the cpu fan BLOW on the heatsink/cpu or have the cpu fan SUCK heat from the cpu/heatsink? Also, i have a p3 866eb with hedgehog/delta 38 combo. And would like to know how fast can this chip be overclocked? Thanks.

 

[kingz]

depends on your case....it doesnt matter a whole lot. just test it. most people say blow onto though.....

 

[dew042]

when the fan blows away from the cpu people generally has some sort of shroud to make sure the airflow is touching all of the heatsink, otherwise just the air around the top of the HS gets moved...anyway, don't know if that makes sense, but I would greatly recomend blowing into the cpu...

 

[duke]

Directing the fan toward the heatsink will result in more efficient heat transfer because of the effect of jet impingment.

 

[EvilDonnyboy]

thats all some nice theorectical mumbo jumbo stuff duke. but how bout real world perfomrance. remember, in theory, rimms are supposed to be 2X faster than pc133 sdram.

suck or blow. it depends on how the hs waz designed. alphas with their thin fins & shrouds are designed to have suking fans. Globalwins (in general) with their thicker fins w/o shrouds are designed to have blowing fans.

 

[Supradude]

you mention the 866EB, ... you probably won't get too much from this chip (hate to dissapoint you), but 133 based fsb chip will not OC much more unless you have good ram and an 815 chipset, or a very tolerant BX...

 

[Sunny129]

in reference to OCing: like Supradude said, an intel CPU that defaults at a 133 mHz fsb, has little headroom to OC since the only way to OC an intel CPU is through the fsb. going from 133 fsb to 150 fsb is possible, but unlikely, whereas going from 100 fsb to 133 fsb is much more probable (with a P3 "E" CPU, which defaults at 100 mHz fsb).

what also makes the 866EB a poor OCer is the fact that the Coppermine core seems to top out at around 1GHz (which is one of the reasons the P3 1.13 GHz CPU was recalled - it was put on the market prior to insufficient testing, and they were just pushing the coppermine core to its limits while trying to say it defaults at such a high speed.) not only will an "EB" not OC as well as an "E" due to fsb limitations, but the 866 is already 133 mHz away from the average 1GHz limit of the Coppermine core. some might go to 1.1 GHz, and others might not even make it to 1 GHz. its all chance, but statistics definitely show that 700 mHz tends to be a sweet spot for OCing, and as CPU speeds exceed this speed more and more, their overclocking probablility decreases dramatically. but its like i said, its all about the luck of the CPU.

 

[longhorn]

I saw a study done on a web page about a year ago on
this subject. On a Celeron, using a GlobalWin Fan,
benchtop test, the processor ran about 2 degrees cooler
in the "suck" direction. I saw similar results on
my Celeron 366-->550 on a slotket.

The Durons run hotter, and I have not seen this sort
of testing done on the AMD chips. On my new Duron
600-->950, I did not flip over the fan. I destroyed
my 1st Duron putting this thing together, and now that
it is working, I don't want to mess with it.