(in
Celsius) |
case temp |
slot
fan case temp |
BGA temp
|
slot fan BGA temp
|
total avg. improvement |
no intake fan |
30.7 |
29.1 |
53.8 |
51.4 |
2 |
intake fan |
31.2 |
26.6 |
53.4 |
51.1 |
3.45 |
no front to
case |
- |
- |
53 |
48 |
5 |
We
used Quake III
to bring the temperature of
the BGA chipset on the AGP card up a bit. Curiously, the cooling factor was lower
then what we expected. The addition of an intake fan helped things out a
bit and gave us a ~3.5 degree overall temperature improvement. One other reason, which we
tested for, was proximity of the slot fan (and thus airflow) to the chipset. With
a bare chipset, there is a scant 14mm space, with the fansink on the
TNT2, that distance was halved. A factor which we couldn't test different situations for was
the position of the intake on the slot cooler relative to the BGA chipset. We were able
to test the affect of proximity vs. cooling ability however.
|
In this digital
composite, you can see how things would line up when the slot fan is
next to a video card. The slot fan is reversed though - in
reality the intake would be facing the card.
On the TNT 2 Ultra we used for testing the position of the BGA
chip-set relative to the fan intake was the same as what's show above
(Note that card shown ain't a TNT). The slot fan covers about half the
chipset.
|
Slot cooler position |
case temp |
BGA temp |
Ambient |
next
slot |
29.6 |
51.1 |
22 |
1 slot down |
27.7 |
50 |
22 |
5 slots down |
30.1 |
51.0 |
22 |
(These values are based on use of an intake
fan & the slot cooler)
Moving the slot cooler one slot down from the AGP card
improved cooling performance by a small but significant value. We had Quake III
going for a while and BGA temp was pretty much solid at 50 degrees. We also
moved the slot fan 5 slots away to see what influence it was really having on
the BGA chipset. While chipset temp rose back up to 'next slot levels,' case
temp also rose ~3 degrees.