Where oh where should the little rad go?

[Gooserider]

I have a bit of a dilemma to settle in my new system. I am trying to decide how I should set up my case so as to get the best cooling. I'm using a dual CPU mobo, and several hot hard drives, so my cooling load is fairly heavy. To minimize noise, my plan is to WC all the major heat producers except the PSU, and then cool the rad with a few 120mm fans, since big fans are supposedly quieter than little ones for a given amout of air movement.

I want to keep the case positively pressured to reduce dust buildup.

I've been involved in lots of discussion in other threads about the water flow and so forth, and I think I have the wet side of the setup pretty well figured out. What I'm after now is discussion of the best way to do the air flow, particularly where the rad should go in the scheme.

I want to see if anyone has any comments on any of these schemes.

My case is an AMS CK1100, which is a 'server cube' style case, similiar to, but slightly larger than, the YY cube case Kevin reviewed here: Yeon Yang Case Review I'll be doing a review on this case here shortly.

This case has two chambers, one containing the mobo and three drive bays, the other containing the PSU and the other 13 drive bays. The way the case is set up, it is obviously intended that most of the intake air comes in the front on the mobo side, and then splits so some goes out the back of the mobo chamber and the rest goes out the drive chamber.

Given the heavier than average cooling load, I plan to use a BIG auto heater core. The heater core database, gave me a choice of several cores, all about 6" wide, 10-11" tall and 2 or 2.5" deep. My plan is to setup the case with four 120mm fans, two intake, two exhaust, and use a thermostat control (either DigiDoc or mobo based) to keep most of them off or at very low speed unless things heat up to much. This should give me a very quiet setup with lots of reserve cooling capacity.

My original plan was to hang the core off the back of the case with the exhaust fans blowing through it, and hope the exhaust air wouldn't have been warmed to much on it's trip through the case. But this would have looked clunky, and made an already huge case even bigger. In theory, I could put the core inside the case, and have the fans suck or blow the air out through it, depending on how I set things up, but this would cost me my 5 internal drive bays. (leaving me with *just* 11 externals ) It would also be a bit tight since the drive chamber is narrower than the mobo chamber.

Looking at the case after I got it, however, I may have a different plan.

On the MOBO side of the case, there is about 5" of space between the front of the mobo tray and the metal front of the case. If I cut the front of the case, I have about another 1.5" before I get to the inside of the plastic bezel.

It looks to me like I could make a cutout in the metal, and put the rad (with a peice of AC filter material in front of it for dust control) half in / half out of the front of the case so the bezel just clears it, On the back side of the rad, I can build a plenum to mount two fairly strong 120mm fans, sucking through the rad and blowing into the mobo chamber. (I'll put louver flaps on each, so as to prevent 'air shorts' when one fan is off or running slower than the other) I would then put a 120mm exhaust fan (mounts already present or I would have used a 92mm) in each chamber for use if needed.

This has the advantage of giving the rad the coolest air possible, but the downside of pre-heating the air going through the rest of the case. I might also have problems keeping the case pressure positive if the exhaust fans both kick on. (Though I don't regard this as a real big issue)

I could reverse the sequence and put the fans on the outside blowing into the rad, but my understanding of how best to do fan airflow suggests that this would be a bit tight in the available space, especially if I want to do an air filter.

What do people think?

Gooserider

[Gooserider]

Let me try something simpler...

Two questions:

1. Assuming the rad and fan are connected by a plenum so there is no air leakage, and that the plenum gives at least a couple of inches between them (to avoid surface interference problems)

Is it better for airflow to have the fan blowing air into the rad, or pulling the air through it? Or does it make a difference...

If one is trying to keep the noise down, will having the rad between the fan and the outside reduce the noise?

2. Assuming that there are fans on the intake and exhaust of the case, and air only enters and exits the case through them...

Is it better for cooling to position the rad at the intake of a case, so that it gets the coolest air, (but then blows that slightly warmer air through the case) or at the exhaust of the case, where it sees warmer air but then gets it out of the case?

(I know the otpimum approach is to pull outside air through the rad and then vent it to the outside without going through the rest of the case; but that would then require additional fans for case air circulation, which IMHO is not an acceptable option.)

Thanks,
Gooserider

[bigben2k]

Have you considered mounting the core on the (non-mobo) side, and ducting air through one or two drive bays? It would give you a fresh intake at the front, and an outlet at the back. You can then ventilate the rest of the case any way you see fit.

I'll be using a blower in my Chieftec server case, and I'm opting to setup an intake on the side (with filter), and outlet the back, where the core (~6 by 8) will be. Anything else will have a set of small quiet fans, for positive pressure.

[Gooserider]

bigben2k Have you considered mounting the core on the (non-mobo) side, and ducting air through one or two drive bays? ...

Yes and no... The way the case is set up, I'm pretty much limited to either the rear at the top of the case on the PSU side (exhaust), or the lower front on the mobo side (intake) Anything else would require lots of surgery, (above and beyond what those two locations require) and / or use of a smaller rad.

I'm also working from the assumption that more fans = more noise, even with quiet fans. Thus I don't want to have seperate sets of fans for cooling the rad and cooling the case if that means more fans total.

Now as I understand some of the other threads, if I get halfway reasonable air flow through the rad, (which is also getting good coolant flow from the high-volume system design) then the air is only going to pick up 1-2 deg. C going through the rad. (remember I'm not doing pelts or other radical cooling, and won't be OC'ing) If I put the rad on the case intake, that means the air going through the case would be only slightly above ambient.

If true, I doubt that 1-2* will make all that much difference in cooling the stuff in the case that isn't WC'd - If there is anything else that gets overly hot, I can always slap another block on it. My pump is rated to do 8gpm into a 14' head, and it's P-Q curve says the volume stays high until the restriction gets over ~10', so I have capacity to spare...)

While the stock case has lots of spaces for additional fans, it only comes with 3 - 80mm intakes on the mobo side and 1 - 120mm exhaust (plus the PSU) on the PSU side. Even sucking through the restriction of an air filter and the rad, I suspect I'd have more volume with the 2 120mm intake fans I proposed in my first message of the thread. I'd maintain positive pressure by either running the exhaust fans slower, or just leaving them off most of the time.

I keep thinking it's almost like a car... The rad on a car warms the air considerably over ambient, but they STILL prefer to run that air over the rest of the engine compartment instead of trying to do seperate flows....

I've basically reached the point of saying that unless someone comes up with a specific reason why I shouldn't, I am going to go with the front mounted rad set up so the intake fans suck through it. It may not be the perfect design from an absolute cooling standpoint, but it is arguably the best use of space, and I suspect it will be only marginally less effective.


Gooserider

[bigben2k]

Either way...

I'd consider ducting (a large duct) from the front, using 1 or 2 bays, since there's plenty of space.

[MadHacker]

I'm in the same boat.
Getting a YY-Cube and had to figure where to put the rad.
I have decided to put it in the rear on the drive Bay side.

Putting a rad above the power supply didn't sound like a good idea. If I do get a leak... don't want it dripping on the power supply.
So i decided I will move the powersupply to the top, and the Rad on the bottom. As for ducting, I thought i'd duct it from the side.
Make a composite shroud (thank you #Rotor for the how to) that is curved to draw air from the side and pass it out through the back.
This system would use 2 120mm fans to cool of the rad. the core i plan to use is the same one Ben2k is using:
core#2-304. It seems to be the best fit in the case.

I would still have one 120mm fan blowing on the Mobo side of the case. If I use aluminum frame fans like
this one it is 80CFM at 30db. I imagine it will be very quite running at 7 volts.

This idea my help..

[Gooserider]

bigben2k: Either way... I'd consider ducting (a large duct) from the front, using 1 or 2 bays, since there's plenty of space.

Well I still don't see why I need to sacrifice bays, when there is a HUGE space on the front of the mobo side that seems like it's begging for a rad. The mobo chamber on the YY cube is fairly narrow, just wide enough for the mobo and some full height cards. The AMS CK1100 case is MUCH wider (16.5"), mostly on the mobo side. The CK1100 is also longer (21") than the YY case, with almost a full length 5.25" bay in front of the mobo tray.

Thus, under the 2 5.25" bays on the mobo side is a space about 11.5" high, almost 9" wide, and about 5" deep with nothing in it but 3 80mm intake fans and the PC speaker. If I make a big enough hole, I can slide the rad out the front of the metal case body about 1.5" (30-40mm) and still leave it inside the plastic front bezel.

The downside is that this space is the only place on the front of the case forced air intakes can be put unless one uses drive bays to mount fans, which is possible but awkward. The back of the case has mounts for two 120mm fans (one in each chamber) plus the PSU and it's fan.

I am going to use a BIG rad. Until I found that the heatercore database listings don't include the tanks, I was thinking of the 2-206 or equivalent (10.75"x 6.125" x 2") However If I have to include tanks, then that one would be a bit to tall, and I will have to find something shorter but wider. I have seen something about the tanks adding about 2" to the listed height, which means I need something about 9-9.5" tall, and around 8" wide.

If I ducted the rad out the side of the case, I would need to either add additional blowhole intakes in the side of the case (which would stick into the mobo space, and mean additional fans / noise) or let the case run at negative pressure sucking in through the stock vents on the sides (which I was planning to cover with sound insulating material).

IF the rad will only warm the intake air a degree or two over ambient(and nobody has said differently) then I don't think it is a big issue to use that same air to pressurize the case, and save on fans at the same time.... I'd put the fans in a rotor style plenum so they blew over the mobo. If the mobo side of the case is pressured, air will naturally blow into the PSU side, and out the PSU fan and the two 120mm exhaust fan holes. If this is sufficient I could probably even leave the exhaust fans off most of the time, further reducing noise.

I can stick the pump on the bottom of the case on the PSU side, just under the PSU. The resevoir and fill tank I can put on the PSU side between the PSU and the drives, with the fill cap sticking out the top of the case, or possibly recessed into it. The net result would be that the only plumbing over vulnerable parts is what goes to them, which is kind of hard to avoid.

Another reason I don't like side intakes and exhausts is that this case will be going on the floor, probably under a desk. I'll want to push it to one side where the side of the desk will tend to obstruct airflow. If I push it to the other side, it will be blowing air on my legs. OTOH, the front will stay clear just because it's the front, and the back is going to have several inches of clearance in order to have room for the cabling.

So far what I've been seeing is what you are going to do, but no reasons NOT to do what I've been proposing....

Gooserider

[shiltz]

i'm using a BIX 2 rad in my current system which is about the same size as the heatercore you said you were gona use and i'm using a YY cube case, what I did that works really well is I mounted the rad flat on the bottom rear of the case on the drive side sucking air from under the case since it stands up on casters and then exausting it out the rear of the case. I did it that way as the idea of sucking air from the front and blowing the hot air into the case effecting the heat on anything not watercooled didn't sound to effective to me nor did mounting it on the back/top and pulling warm air from inside the case sound to effective either.

[Gooserider]

I asked the question about Delta-T going through the rad in another thread - HERE

The answers I got in that thread confirmed my recollection that air sucked through a radiator will only go up a maximum of 1-2 degrees. If one has decent airflow, the Delta-T will probably be well under a degree.

Now it might be possible that raising the incoming air temperature a couple of degrees above ambient could screw things up, but I have a hard time believing it.

I don't know just where the notion that the air coming off a WC rad is so hot that it must be kept out of the rest of the case came from, but it appears wrong to me.

I did see several comments about being sure to duct the rad air out of the case in some of Joe's early articles on the site. But these articles date back to the days when, by his own admission, he knew alot less on the subject, and may not have had the numbers on how little the temps really changed. (plus many of his boxes were running pelts, doing heavy OCs or had other cooling loads which gave him a bigger Delta T)

I think many people might also see 'heater core' and start thinking in terms of their car heater experiences; forgetting that there are much higher temps in an automobile than they are in a WC circuit. I know that was my initial reaction, and I was surprised when I found the number was as low as it is.

I don't see ANY other advantage to ducting the rad out of the case, and considerable disadvantage. So why NOT use it for cooling? Bear in mind that I'm going to be WCing all the major heat producers in the case, so I shouldn't get the big increase in temp on the way through the case that the air cooled guys do. Frankly, I'll be surprised to see as much as a 10* increase from the case intake to the case exhaust. (except the PSU, which is exhausting nearly all of it's heat outside the case.

shiltz: i'm using a BIX 2 rad in my current system which is about the same size as the heatercore you said you were gona use and i'm using a YY cube case, what I did that works really well is I mounted the rad flat on the bottom rear of the case on the drive side sucking air from under the case since it stands up on casters and then exausting it out the rear of the case. I did it that way as the idea of sucking air from the front and blowing the hot air into the case effecting the heat on anything not watercooled didn't sound to effective to me nor did mounting it on the back/top and pulling warm air from inside the case sound to effective either.

See my note on the temperature change, and ask why it's a worry...

However your idea doesn't sound bad except that I don't like the flow direction you're using. It reminds me of a tale I heard several years back about some main-frame designers debating airflow direction. There were several 'pull from the bottom' advocates who changed their views when one opposed to the idea mentioned his broken english conversation with a new immigrant cleaning lady who told him how easy it was to clean the computer room because the dirt 'just disappear' when she swept it under the machines...

It might be OK to pull from the underside IF one is using (and regularly cleaning!) a good air filter, but I think you'd have better results to reverse the fan so you are sucking from the back and blowing through the rad out the bottom. Even doing that, I'd still use a filter. Remember that a rad full of dust bunnies isn't going to cool very well! (and rads are hard to clean...)

I did consider doing an exhaust duct out the bottom, which would be possible in my case, and appeals to me more than a side duct. I decided against it because it would still leave me with the problem of having to add additional intake fans and finding room for them.

Gooserider