which one is a better idea for rad placement?

[Rayman2k2]

I am planning on building an external box for my radiator, and I had a couple questions. First, should it be 1 radiator with fans blowing in. Second, should it be 2 radiators with 2 120mm fans blowing in, but not "shrouded" onto the radiators? The box has to be 6" wide, no more. So with 2 2" thick radiators, it would be hard to stick a 2" shroud on to it. The radiators are 9" by 6" by 2", and will be placed horizontally (so the 9" side is on the bottom). Also, what if I had two radiators with two fans in between, pulling air through one, and pushing it through the other?

[Gooserider]

One rad should be plenty unless you are doing some really big time heat generation.

It is better to suck through the rad than to blow into it. However it is also important to use some sort of shroud to ensure that all the air goes through the rad instead of bypassing it. A shroud should also help space the fan at least one inch away from the radiator surface.

Note that there isn't anything 'magic' about the shroud shape (other than avoiding restrictive designs) and you may be able to design your box so that the box material functions as a shroud w/out extra work.

Gooserider

[Rayman2k2]

well here are some of my ideas:

^^^^ <-Airflow
 | | | |  <-Airflow
====  <-120mm Fan
[      ]   <-Rad
[      ]    <-Rad
-------    <-bottom (no gap)

The airflow will be forced to go through both rads

R\____/R
R   F  R
R __F_ R
R/    \R

The air will be pulled through one rad and will have no choice but to be pushed through the other due to the shroud.

R = Rad
F= Fan

[PooJou]

Ok, just get a box, cut 2 X 120mm holes at once end, get one big arse rad or smiliar, cut a hole, and mount it up, sealing the edges, put pump etc. inside box...

Shut the lid :P - air will be pulled through the rad as thats the only place the fans can get air from, though if you're making a big external box, using Bigger beefier (those 170mm Comair Rotrons for example) to pull air would be beautiful :P

[superart]

Raymann2k2, not to knock anyone’s ASCII art, but in the interest of clarity, next time, could you just do a quick sketch in paint and attach that instead of making your schematics using ASCII characters.

Thanks.

[Rayman2k2]

Originally posted by superart
Raymann2k2, not to knock anyone’s ASCII art, but in the interest of clarity, next time, could you just do a quick sketch in paint and attach that instead of making your schematics using ASCII characters.

Thanks.

lol sorry!


the first setup is just what PooJou described, only with two radiators.

the second setup has a radiator with a shroud, with two fans pulling air through the radiator. The air is then forced through another radiator (with the help of a shroud)


but you got to admit, ASCII art = teh shit!!!!1

[prb123]

Why would it be better to suck air through the radiator? I thought that blowing air is better due to it being more turbulent from passing through the fan(s).

[Gooserider]

I still don't see why you are insisting on two rads, there is little or no benefit to dual rads in terms of cooling unless you are doing several independent loops (a marginal benefit choice unless using TEC's) or are doing something really severe in terms of cooling loads (like MULTIPLE PC's)

Dual rads will greatly increase your liquid flow resistance (if only because you are doubling the passage length) and totally hose you on airflow resistance. A single rad of reasonable size should be more than enough for any reasonable cooling load.

As to the reason for sucking being better than blowing, my understanding is that when you are sucking the fan creates a partial vacuum on one side of the rad and the ambient air pressure effectively pushes the air through the rad to fill the vaccuum. When blowing, you are trying to create pressure that comes through the rad and displaces the air that is already there. This is less efficient. Also sucking tends to be easier to control the air flow so that it all goes through the rad.

Turbulence is a mixed bag - it restricts flow by definition, which is good for thermal transfer but given the low heat capacity of air, you are better off with more volume. (Part of the reason for using shrouds is to allow the turbulence to smooth out...) Just going through the rad will induce more than enough turbulence to give good heat transfer.

Gooserider

[Althornin]

Originally posted by Gooserider
I still don't see why you are insisting on two rads, there is little or no benefit to dual rads in terms of cooling unless you are doing several independent loops (a marginal benefit choice unless using TEC's) or are doing something really severe in terms of cooling loads (like MULTIPLE PC's)

Dual rads will greatly increase your liquid flow resistance (if only because you are doubling the passage length) and totally hose you on airflow resistance. A single rad of reasonable size should be more than enough for any reasonable cooling load.

Just use the two rads in parallel.
Now, you have lower flow resistance.
More radiator surface area is better. Gets you that much closer to ambient, plus allows room to grow and not have to redesign the system.

[Gooserider]

You are right about less resistance from parallel rads, assuming that the extra plumbing needed to put them together doesn't cancel out the lower resistance.

On the second point, more rad surface is not necessarily a good thing if it causes lower air flows. Thus far the designs I've seen suggested in this thread would have far less total airflow than a single rad design, which will cause WORSE cooling, not better.

Growth is not that big of a problem, assuming you start with a reasonable size rad. The proposed designs featured a 6" x 9" x 2" rad, which is more than sufficient to cool any single PC. This is especially true if using a single pass design core like the FedCo # 2-342, which is about that size. Assuming an actual flow rate = to 50% of the free air rating, and a 10*C MAXIMUM temp increase, then the 2-342 will handle approx a 500W load w/ only two 120mm x 32mm fans with an 80 CFM rating. (and this is a very conservative estimate) By my estimates this would be enough to cover a system with dual CPU's, a hot video card, and a good assortment of NB's, HDD's, etc. plus a high power pump, with plenty of margin.

Remember also that approaching ambient is an asymptotic curve, so doubling the rad area is NOT going to halve your Delta T. Depending on how close you would have been with a single rad setup, doubling the area might only get you a couple of *C, which is hardly worth it.

Gooserider

[Rayman2k2]

well, i thought that dual rads would also help in running the system periodically without fans (quiet cooling), but if one rad can handle that, than, I'll just stick with one!

[Althornin]

Originally posted by Gooserider


Remember also that approaching ambient is an asymptotic curve, so doubling the rad area is NOT going to halve your Delta T. Depending on how close you would have been with a single rad setup, doubling the area might only get you a couple of *C, which is hardly worth it.

Gooserider

never said it wasnt asymptotic.
A couple of C may not be worth it in your eyes, but thats all that seperates the top blocks from one another.

[Gooserider]

Althornin:
quote: "Originally posted by Gooserider Remember also that approaching ambient is an asymptotic curve, so doubling the rad area is NOT going to halve your Delta T. Depending on how close you would have been with a single rad setup, doubling the area might only get you a couple of *C, which is hardly worth it. Gooserider"

never said it wasnt asymptotic. A couple of C may not be worth it in your eyes, but thats all that seperates the top blocks from one another.

True, but I tend to look at things from a cost / benefit standpoint... Buying one block instead of another is relatively low cost (just the difference in price between the two), so picking up a degree by using the better block (and I will say I wonder about the reliability of numbers that close as to which one is 'better') is worth it.

OTOH, the dual rads incurs a great increase in plumbing complexity, possible issues of increased noise, possibly bigger cases or cooling boxes, etc. IMHO the cost is not worth the extra headaches; especially since I may have overestimated the benefits.
-------------------------------------------------

Rayman2k2 well, i thought that dual rads would also help in running the system periodically without fans (quiet cooling), but if one rad can handle that, than, I'll just stick with one!

A larger (or multiple) rad WILL help enable passive cooling, but only if it is setup so that there is good natural convection airflow through the rad Putting the rad in a box is probably NOT going to give you this. The passive units I've seen pictured tended to use automobile rad size units, mounted externally so as to maximise the passive airrflow.

Gooserider

[georgeteo]

Building a Rad box is no different from sticking the Rad inside your case.

If you want maximum performace, its best to leave your rad outside the case in open air.

Multiple rads would improve the cooling performance of your system. But at what cost?? money, loss of flow rate, space constraints?? Fans play the major role in your Rads performace. Increasing fan speed, no. of fans, fan placement, shrouds are just some of the ways. If its performace your looking for, I'd suggest getting the thermochill 120.2 or 120.3 with 4 or 6 deltas in push pull config.

The only reason I can think of having dual rads is if you are looking for performace at a low noise level. Juz look at my Sig. I'm doing fine with heavily o/c system, GPU block and a 700gph pump submerged. My fans?? 4 in push pull at only 5v (My rig is geared towards low noise).

Juz remember to place your rads just after the hottest part of your system inn order to maximize its cooling potential. My rad is placed juz after my pump.