SteveH said:
Not if you have quiet fans to start, and you can regulate the speed.
Not if you have the intake fans filtered, as they are in my PC case (have a
look at a Thermaltake Xaser III)
In my case it has nothing at all to do with defective hardware. The memory
passes extensive testing with Memtest, the HDD's pass their fitness tests as
does any other part that I can reasonably test WITHOUT wasting hours testing
each part down to component level.
You don't listen to others, do you? I stated in my previous reply that I
have built plenty of PC's with only two fans. I even accept the fact that
its perfectly possible to build a PC with no fans, provided the box isn't
full of heat generating extra parts, like mine is.
As I said, with one HDD, one optical, non-overclocked CPU and memory and an
average video card, your one or two fans theory, is perfectly possible, but
in my PC it isn't because of all the extra (heat generating parts) parts.
And I don't need science to tell me what I can clearly see for myself.
But my main point still stands - you can't just genralise on these things
the way you originally did. Every case is different.
Well, actually, you can 'generalize' but you're also correct that each case
is different.
Case cooling is the thermal capacity of the cooling medium (air) and how
much of it to move based on how much temperature rise you want.
So, in basic theory (ignoring lower order altitude and temp variations)
Delta-C = 1.71 * Watts/CFM (for air)
And I would imagine that's the origins of where w_tom is coming up with his
25-30 CFM 80mm fan although, while he constantly pontificates about
'engineering' and 'numbers', he rarely provides anything, in that regard,
of much practical use, like the actual equation.
However, you will notice right off that there are two assumptions in there:
the desired case temp rise and the total power being dissipated in the case.
Things not obvious from the equation are that, for one, and depending on a
host of case factors, you'll likely get less than half the rated CFM from a
case fan in actual, mounted, use. So, if you need 25 CFM, a 25 CFM fan
isn't going to do it and you'd be looking at a 50 CFM fan (if you're doing
positive case pressure filtered intake you can forget getting anywhere near
50% even before taking into account filter clogging over the maintenance
interval). Second, the equation presumes an even temperature across the
case and uniform airflow through it; neither of which will be reality.
To make matters worse, CPU heatsink fans can act as a 'stirrer', in
addition to cooling the heatsink, and dilute the case fan's effect because
the warmer HSF air is constantly remixed with the incoming case airflow.
That effect is minor with a small HSF, as with older low power processors,
but can become significant when the HSF's volume is nearly as large as the
case fan's. Regardless, the air temperature around the processor, and other
'high power' components, is going to be higher than the 'calculation' case
temp even before one considers uneven case airflow distribution.
Another factor missing from the equation is just how much noise you are
willing to live with and (if we're looking for 25 CFM total) two 25 CFM
fans are quieter than a single 50CFM fan of the same size. Or one could
change to a larger diameter fan, IF the case can support a larger diameter fan.
The long and short of it is, while the 'theory' may say 25 CFM is enough,
depending on a number of variable assumptions, reality can be quite different.
In addition to 'high power' component cooling, which is needed simply to
get their heat into the case air, you can end up with hotspots (case
airflow inconsistencies) necessitating additional localized cooling and
more than one exhaust fan may be desirable from a CFM/noise aspect above
and beyond the raw 'horsepower' to get the needed CFM. And intake fans may
be desirable to direct inlet airflow to a particular portion of the frontal
area, such as over hard drives.
And then there's the fact that it's easier to approach 'theoretical' when
one can design their own case but when putting fans in an existing case
you're often stuck with whatever 'inefficiencies' the case designer built
into it and adding a second exhaust fan is more practical than trying to
reverse engineer case airflow directors/ducting.
So, 'everybody' is 'right'. In theory, given complete design freedom and no
restricting criteria, such as noise, fan size, etc., one can cool a
'typical' PC with one case fan. However, give the particulars of a specific
PC it may be quite desirable to have more than one.
Having said that, there does seem to be a tendency for some home builders
to 'over fan' their cases for little benefit.
