In message <
[email protected]> kony
It tends to be a bit of an issue here, largely because of all the junk
near the top of my case. I have two optical drives at the top, then a
removeable IDE drive below -- The removeable drive sticks further into
the case then my optical drives.
As a result it forms a "shelf" that catches the ends of the spare cables
and it ends up being impossible to remove the IDE cables going to the
optical drives when I need to pull out an optical drive because of the
power supply cables being in the way.
I've got them on their own IDE cables right now, although I'm not sure
it's totally required anymore since they're both UDMA (Mode 4 and Mode
2) -- When I originally built this system I had one CDR that wasn't
UDMA. The extra IDE cables don't help either.
Your call... if the drives won't be used simultaneously you
probably don't "need" one cable per, but if you can get it
to work I'd do so. Optical drives don't need much airflow
though, overlapping ribbon cables isn't too bad, sometimes
even easier than dealing with round cables that are too long
(as is often the case since cables long enough to reach the
top of a mid/ful tower case may not have connectors spaced
very well). Seems like cable manufacturers ought to make
more cables with both connectors only 3" apart from each
other since that's often how far apart the drives are.
I don't always have that many drives, but my file server just died and
most of the drives were moved into my web server, but 2 of the IDE
drives don't fit (My web server is out of IDE ports, PCI slots, and
power connections, and has neither USB2 nor Firewire, so it's not a
simple matter to install the drives -- I'll just toss them into my
machine in the mean time, they're mostly used for backups of other PCs)
I'm actually starting to overload my current supply with only 2 IDE
drives (one in the removable bay, one mounted internally) -- It's a
decent supply, but it's not up to what I want to do with it
In truth my case is too small and I need to try and find a larger one,
but I'm having a lot of trouble finding something reasonably priced that
meets my specs.
That's part of why I got into moddin' older cases. After
adding another drive bay, one of my full towers holds about
15 drives +-2. It cost about $10 plus some fabrication time
for the added bay... and was big enough to hold two power
supplies. Boreing beige color though, but I've already
outgrown "uber-elite" looking cases, have fancy ones and
after a while didn't care so long as the drives aren't the
wrong color... which a can of vinyl dye will often fix, does
very well for painting plastic case parts and optical drive
faceplates, etc. Even so, mostly I use it for other
people's (kid's) systems.
One of the things I really like about the Neopower 480 is that it has a
completely separate 12V rails, so the motherboard and CPU are running on
their own 12V rail and don't float as the load changes.
That is good, providing the split rails each provide ample
capacity. Suppose you have a ton of drives though and each
rail is spec'd for 14A... you might want more reserve on one
of the rails for drives spinning up at power on, and not
need that many amps for CPU initialization. In that case it
might've been as well to have a single 28A 12V rail.
Even so, most PSU I've seen with the split rail design do it
a certain way one "could" implement themselves if they were
motivated enough to do it. On each of the 12V rails there's
usually a very low-ohm resistor, so two low-ohm resistors is
what isolates the two rails opposed to none on the single
rail design. It means the PSU has accounted for this in the
voltage feedback/control circuit too AFAIK,but point being
it isn't necessarily a substitue for raw amperage capacity,
a beefier PSU IF that's possible.
I've yet to see the voltages go beyond +/-2% whereas my old supply would
regularly float +/-9% under high load.
9% is pretty bad, but is this from a multimeter reading or
motherboard sensors?
I'm actually running the Neopower 480w right now, but I find people tend
to be less critical of hardware if I say I've already purchased the
hardware vs a prospective purchase -- I'd rather hear I ****ed up and
what I bought is shit (if that's the case) then have people make nice.
IMO, the real issue is how much power density (output) a PS2
form-factor PSU can support with both A) Long life B)
Low noise
For this reason, my main gaming rig has two PSU in it, as
does my file server. I buy PSU at good spot-pricing though,
it may not have cost me much if any more to do this than
some pay for a single PSU, so all things must be put in
context.
So far I've yet to see the power supply fan go above 900rpm (according
to my motherboard -- The Neopower power supply has a connector that
allows the motherboard to monitor the speed of the power supply fan) and
my case temperature (again according to the motherboard) is around 25C.
That's quite good, but what's the room temp?
Typically a well-engineered quiet system (with fair amount
of modern heat-generating gear) will be at least 5-10C above
room ambient if the fans are low enough to be quiet. It
does depend where one takes the temp reading though. I
mostly care about HDD temps and capacitor temps (on video,
mobo, and PSU), except when overclocking... then whatever is
applicable to the situation.
FWIW, my case temperature sensor seems to be fairly accurate, or at
least it matches the thermometer I put in my case when I built this
system. The ratings were initially WAY off, but a BIOS upgrade seems to
have helped.
BIOS temps are not case temps. They're one single chip's
integral sensor more often than not, and even fluctuate a
lot based on a heatsink (or lack thereof) on that chip, as
well as a front fan... a front fan can make a bios system
temp reading much lower while not necessarily cooling whole
system as well as it cools the drives. Drive cooling is
important to me, I value the data more than the hardware
quite often, but backups take away a lot of the concern.
One of these days I'll break down and buy a proper temperature monitor,
but until then, the motherboard isn't too far off the case's air
temperature, so I'll use that as a guideline.
There are lots of ways to do it, if you don't need
*constant* temp feedback you might consider getting a
multimeter with a temp sensor function, then test system at
full load in high ambient room temp to get a maximal
reading... then all you have to do is adhere to regular
cleaning schedule as needed, and monitor fans so be sure
there aren't significant failures... though that's one of
the nice things about larger systems, often a single-point
fan failure won't bring whole system down, except maybe
video or CPU depending on the parts.
I have two fans at the bottom/front of my case which blow directly over
my 2 Raptor drives, these two run at full speed at all times. IIRC one
is intake the other is exhaust, although I'd have to pop open the case
to check for certain. When I got this case I spent a while tweaking the
fans direction and positions to see what resulted in the coolest drives
and CPU temperatures, and I'm pretty sure that an intake+exhaust right
at the Raptor drives turned out to be best.
I also have a side intake and rear intake which is regulated by the
power supply as well and are running at around 900rpm or so)
I modified a PCI-slot VGA cooler to function as an exhaust rather then
an intake and it sits directly above my video card -- It doesn't push a
lot of volume of air, but it dropped my case temperature a few degrees
on it's own when I installed it.
I thought most of them were exhausts? I don't usually use
one of those, rather opting for positve OR negative case
pressurization (depending on if the case if filtered or not)
then I leave that PCI slot bracket off. So far it's worked
well enough IMO.
At the top of my case I have the power supply's 120mm exhaust fan (which
sucks air in from the bottom of the supply rather then the front)
All in all I'm fairly happy with the cooling -- It's positive pressure,
which is mandatory because I have hardwood floors and four cats in the
house -- Positive pressure allows me to place grills over the intake
fans that catch most of the cat hair before it enters the system. My
system turned into a nightmare when I first got the cats, but once I
switched to positive pressure my system has never been cleaner.
Yes cats are evil.
At least they're fluffy evil instad of lizard skinned evil.
It would be kind of cool to have an iguana living on top of
a system though, setting it up so water-cooling heats a
hot-rock on top.
They (the fan grills) get cleaned a couple times a week in most cases,
if I don't, the case temperature goes up pretty quickly.
Yep, that's the problem with high-density filters. My next
build will filter with fine-weaved 3M-like furnace filters
and have a slot on the side to just slide out the filter and
throw it away. Local hardware store has generic versions of
the 3M filters for about $4, one of larger sizes would make
4-6 filters large enough for the entire front intake.
The biggest thing I like is that I can unplug the PCI-E connector, and
that I can (down the road) swap in more SATA connectors rather then
needing additional adapters attached to the regular AT hard drive
connectors.
I'm expecting to be using all of the modular plugs on the power supply,
although that's not certain yet.
I still hope for a more robust SATA connector. I can accept
that it might cost more, and that 30 cents is a cost I'm
willing to bear.
