Advantages of a Dual 120mm Radiator or a Single Fan?

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
whitebloodcell
Posts: 11
Joined: Sun Feb 27, 2005 5:11 am
Location: England

Advantages of a Dual 120mm Radiator or a Single Fan?

Post by whitebloodcell » Sun Feb 27, 2005 5:26 am

Hi all,

I am hoping my Water Cool my new system, my first time water-cooling so am unsure as to what is necessary.

I am going to be cooling overclocked Athlon 64 3000+, 2x 6800GTs Overclocked, And Asus A8N Sli Deluxe Chipset,

Using DD Maze 4 GPU Blocks, DD 12V Pump, Swiftech CPU Block, DD Asus Chipset Block, 1/2ID Primoflex Tubing,

I am am deciding whether a dual 120mm Radiator is necessary. If so I was going to go with Black Ice Pro II Rad. If not, then Black Ice Pro Single Fan, Will be using Nexus 120mm Fans to Cool Rad, probably at 7V.

Are the thermal benefits of a Dual Radiator worth the extra Noise?

Thanks

Chris`I
Posts: 27
Joined: Wed Jan 05, 2005 6:00 am

Post by Chris`I » Sun Feb 27, 2005 5:48 am

From my experience I would say yes.

I run an XP2400+, R9800pro and NB all used to be cooled by a single BIX and more recently 2x BIX. The double rad combo was much better at cooling dropping the temps by a few degrees. With your system, you will be producing much more heat so I would see a dual rad as essential! If nothing more it will allow you to run 2x 120mm fan at a lower speed than a single one.

The system components I run are pretty similar. I have DD Maze 3 CPU, Maze 4 GFX and NB cooler. Running off a Ehiem 1048

whitebloodcell
Posts: 11
Joined: Sun Feb 27, 2005 5:11 am
Location: England

Post by whitebloodcell » Sun Feb 27, 2005 6:02 am

Is Mounting the Radiator a difficult Job? I am hoping to mount it in the upcoming Antec P180, or If I can't wait, the P160? Do the Fans attach directly to the Rad, or do you sort of just position them in front of it? Is it more efficient/quieter to pull air through the radiator? So you mount the fans in the Top of the Case, then somehow suspend the radiator beneath it? Or attach the Rad to the Case and Blow Air through it?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sun Feb 27, 2005 7:45 am

whitebloodcell wrote:Is Mounting the Radiator a difficult Job? I am hoping to mount it in the upcoming Antec P180, or If I can't wait, the P160? Do the Fans attach directly to the Rad, or do you sort of just position them in front of it? Is it more efficient/quieter to pull air through the radiator? So you mount the fans in the Top of the Case, then somehow suspend the radiator beneath it? Or attach the Rad to the Case and Blow Air through it?
Always use a shroud and always install the fan to pull rather than push through. CoolingWorks MiniCoolShroud is a premade shroud designed specifically for usage with 120mm Black Ice Pro/Xtreme radiators and 25mm thick fans (the included bolts are too short for 38mm fans). Another solution is to fab out shrouds by hand if you're looking to save money, but I like how the MiniCoolShrouds come with shroud, gaskets and all necessary hardware.

Do you plan to use a reservoir in your system at all? Or will you be using a T-line to fill & bleed your loop? The reason I ask is because the fillport must be at the point of highest elevation in the cooling loop to get a proper fill and bleedout of air in the loop, and if the radiator is at the top of the loop with fittings pointed straight down, I think life will be difficult indeed. The solution, however, is that because the top opening is out back on P180, you may be able to pull it off if you tilt the entire system back when filling so as to get the top of the T-line or the fillport in your reservoir higher than the radiator's fittings. This seems like a highly inconvenient setup though, given the fact that you still need to let the system tilt back during normal operation and this will affect...things...

Sorry, I'm unsure if the top mounted fan has regular screw holes. If it does, that would make things a bit easier, but I don't think so, given the way the top panel is made (aluminum->plastic-aluminum sandwich, sort of thick).

-Ed

whitebloodcell
Posts: 11
Joined: Sun Feb 27, 2005 5:11 am
Location: England

Post by whitebloodcell » Sun Feb 27, 2005 8:33 am

I was going to go with a Reservoir, I just havent been able to find one that I happy with. Ideally I would like one of the Aqua Cool Ones but I can't afford it, I would have liked a 5.25" Mounted one but I kept reading that they were arkward to fill and mount and many had problems with leaking.

So, irregardless of my impossible wants what reservoir would you reccomed, is there a particular most people go with, I am looking for one that is cheap and does the job well, nothing more really.

Do you agree with Chris said, a dual fan radiator is more suited to my situation? Do Mini Cool make a Shroud for Dual Rads?

What other Components/accessories do I need? I keep reading about T Juntion type things which (when used for an Sli setup) split the water so it goes to bothe graphics cards in parallel. Would 2x 120mm Black Ice Pros give the same cooling as 1x 240mm Black Ice Pro?

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sun Feb 27, 2005 12:55 pm

whitebloodcell wrote:I was going to go with a Reservoir, I just havent been able to find one that I happy with. Ideally I would like one of the Aqua Cool Ones but I can't afford it, I would have liked a 5.25" Mounted one but I kept reading that they were arkward to fill and mount and many had problems with leaking.

So, irregardless of my impossible wants what reservoir would you reccomed, is there a particular most people go with, I am looking for one that is cheap and does the job well, nothing more really.

Do you agree with Chris said, a dual fan radiator is more suited to my situation? Do Mini Cool make a Shroud for Dual Rads?

What other Components/accessories do I need? I keep reading about T Juntion type things which (when used for an Sli setup) split the water so it goes to bothe graphics cards in parallel. Would 2x 120mm Black Ice Pros give the same cooling as 1x 240mm Black Ice Pro?
I use the Swiftech reservoir; the white plastic one with the funny shape. It works very well, just make sure you seal the fittings with silicone or teflon paste. Teflon tape didn't do the job for me. If you use one of these, make sure to keep it as full as possible, or it has a tendency to, "gurgle." In the end, not even teflon paste allowed me to stop the leaking with my clear plexi reservoirs, which I tried two different designs of.

I am not 100% sure if dual MiniCoolShrouds will fit a BIP II; I use two separate 120mm Black Ice rads, one exhausting out the front and one exhausting out the back, with an opening in the side as intake. Dual 120s will definitely be more effective than a single one of twice the thickness. I recommend avoiding splitting the flows to the two GPU blocks so they're parallel; keep them in serial flow, as the temperature difference between the two blocks will be negligible, and not having y-split then y-combine will reduce overall flow restriction compared to running straight serial, plus the actual flow rate through both blocks will be full, rather than half, which will improve thermal efficiency with Maze-type designes like the GPU blocks you'll be using (they rely on higher flowrate to perform optimally).

-Ed

whitebloodcell
Posts: 11
Joined: Sun Feb 27, 2005 5:11 am
Location: England

Post by whitebloodcell » Sun Feb 27, 2005 2:05 pm

Would two Black Ice Pros be as effective as a Black Ice Pro II? Is Flow Rate slowed when going through the Rads signifacntly? Do you have the Rads one after the other in your loop? Two Single Rads would suit me better, Means I wouldnt need to start boring holes in my case, only downside is the cost :( I'm on a tight budget. With the Black Ice Pros, do you need strong airflow going over them for effective cooling? I obviously want as little noise as Possible. Would two Fans at 5V on two separate Rads? Be more effective and quieter than one fan at 12V on one Radiator?

It looks like two shrouds would fit on, do you use the holes around the side of the radiator to mount it? On the Black Ice Pro II there are 4 Holes around each half.
Image

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sun Feb 27, 2005 3:16 pm

whitebloodcell wrote:Would two Black Ice Pros be as effective as a Black Ice Pro II? Is Flow Rate slowed when going through the Rads signifacntly? Do you have the Rads one after the other in your loop? Two Single Rads would suit me better, Means I wouldnt need to start boring holes in my case, only downside is the cost :( I'm on a tight budget. With the Black Ice Pros, do you need strong airflow going over them for effective cooling? I obviously want as little noise as Possible. Would two Fans at 5V on two separate Rads? Be more effective and quieter than one fan at 12V on one Radiator?

It looks like two shrouds would fit on, do you use the holes around the side of the radiator to mount it? On the Black Ice Pro II there are 4 Holes around each half.
Image
I am not sure which situation would perform or flow better, single II or two Is. I have my loop set up as reservoir->pump->radiator1->GPU->CPU->radiator2->reservoir, and the reasoning was primarily that with the way the components are arranged in my system, this setup allowed me the least restrictive tube runs and easiest bleed-out. I've run through several evolutions of cooling loops in my system; here's the thread about it, including more photos, text and test data than you'd probably ever want to see. If you want to see only the most recent stuff, read the thread backwards.

I agree that it looks like two shrouds would work fine on a single Pro II.

-Ed

autoboy
Posts: 1008
Joined: Fri Dec 10, 2004 8:10 pm
Location: San Jose, California

Post by autoboy » Mon Mar 07, 2005 11:49 pm

I am a thermo engineer. Two radiators in parallel work better than two in series because heat transfer works better with a higher temperature difference. Both radiators get water at the same high temperature so they can cool more effectively despite half the flowrate. With two radiators in series the water is cooled by the first and the second gets cooler water which decreases its efficiency. The double fan radiators are not really either in series or in parallel because hot water flows through one side of the radiator down to the second fan and then cool water flows back up to the first fan. However, the two radiators in parallel decreases the head (back pressure) on the system so flow rate increases with two seperate radiators in parallel resulting in greater cooling capacity. However, the effect it will have on your watercooling system will be tiny. Go with whatever is easier for you just make sure you hook em up in parallel. Same goes for the video cards as they are just a radiator in reverse.

bobkoure
Posts: 191
Joined: Sun Aug 11, 2002 3:26 pm

Post by bobkoure » Tue Mar 08, 2005 1:07 pm

autoboy wrote:I am a thermo engineer. Two radiators in parallel work better than two in series because heat transfer works better with a higher temperature difference.
How does having rads in parallel present more frontal area at a higher delta-temp? IMHO it just moves the slightly-higher-efficiency frontal area from one rad to one side of each of the two in parallel. However, rads in parallel do reduce loop resistance, which may pay off in a better C/W in your waterblock(s) (higher flow, more pressure = better C/W, all be it not dramatically so with blocks like the Swiftech 6000). Do remember that the delta-temp is much smaller than what you're used to engineering (I/C from your handle), and the delta between radiator coolant in and out is also pretty small.

As far as slower flow in a radiator being inherently a good thing, the relationship between flow rate and coolant-to-metal heat transfer appears to be pretty much radiator idiosyncratic, so I'd be skeptical of any blanket statements about this that don't have decent measurement data about a specific rad (got data?). As a thermo engineer, you're not falling into that old fallacy of "the water needs to slow down to transfer heat", right?

As far as putting any kind of water blocks in parallel, you're reducing water flow through each block, which may reduce the block's c/w to the point that you're better off running serial, with the slightly warmer coolant (on the second block) but at full volume/pressure, so IMHO you at least need to know which blocks and what flow. It's also possible that you get the reverse effect (better flow in parallel) with the low-noise low-power pumps popular here and very restrictive blocks. Jury's still out.

Note that all this is worth thinking about as a major noise producer is the fan(s) moving air through the rad(s). A better C/W means you can turn the fans down a bit more, let the coolant get warmer, and still have acceptable CPU/GPU temps.

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Tue Mar 08, 2005 1:18 pm

autoboy wrote:I am a thermo engineer. Two radiators in parallel work better than two in series because heat transfer works better with a higher temperature difference. Both radiators get water at the same high temperature so they can cool more effectively despite half the flowrate. With two radiators in series the water is cooled by the first and the second gets cooler water which decreases its efficiency. The double fan radiators are not really either in series or in parallel because hot water flows through one side of the radiator down to the second fan and then cool water flows back up to the first fan. However, the two radiators in parallel decreases the head (back pressure) on the system so flow rate increases with two seperate radiators in parallel resulting in greater cooling capacity. However, the effect it will have on your watercooling system will be tiny. Go with whatever is easier for you just make sure you hook em up in parallel. Same goes for the video cards as they are just a radiator in reverse.
My friend, if you're an engineer, why aren't you realizing the fact that the temperature of the coolant in the loop varies by well under 1 degree celcius between outlet from the reservoir and inlet at the reservoir? This coolant isn't moving at molasses pace--it's flowing rapidly enough that the order of components in the loop doesn't even matter from a thermal standpoint (other issues are far more important in terms of determining order of components in the loop, like reducing backpressure at the pump inlet, preventing tight bends and making the system easy to fill & bleed). It's quite commonly known that the the temperature of coolant barely varies throughout the loop--the coolant in the second radiator will be minutely warmer than the coolant in the first radiator, unless you've got glacial flow and/or an extremely lengthy piece of the cooling loop separating the radiators (try a couple dozen waterblocks with heat sources attached and about a couple hundred feet of tubing to sustain it all). PC cooling loops are (usually) so short that the temperature variance between start of loop and end of loop is absolutely minimal, unless your pump is failing and/or undervolted to the point of partial stall.

If you're talking about having two radiators in serial air-flow, then you're right; two radiators in serial airflow is absolutely horrible. The second radiator, trying to use the already preheated air from the first radiator, will be highly ineffective. In this respect, the difference cannot be understated: two radiators in parallel air flow will perform significantly better than two radiators in serial air flow, particularly with lower air flow rates, such as that which we SPCR members tend towards.

-Ed

Twill
Posts: 34
Joined: Sun Apr 20, 2003 4:05 pm

Post by Twill » Fri Mar 11, 2005 5:01 pm

The reason I ask is because the fillport must be at the point of highest elevation in the cooling loop to get a proper fill and bleedout of air in the loop, and if the radiator is at the top of the loop with fittings pointed straight down, I think life will be difficult indeed.
Hey, Ed, do you have a link with more info on this...and do you know if it has to be permanantly above everything else or only when filling and bleeding??

thanks

Twill

bobkoure
Posts: 191
Joined: Sun Aug 11, 2002 3:26 pm

Post by bobkoure » Fri Mar 11, 2005 6:15 pm

Actually, this isn't my experience at all. I've built a number of systems using Eheim 1046, 1048, and HPPS pumps with the Innovatek agb-o-matic push-on reservoir. The fill port is the top of the reservoir, but I've never had one of these at the top of the loop - and never have had a problem, either with bleeding air (seems to be more a matter of giving the water a point to slow down as otherwise entrainment is just pulling the bubbles along) or with water coming out with the system stopped and the fill port open.
Maybe we're both misunderstanding what Ed's saying (writing)...

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Fri Mar 11, 2005 6:23 pm

Twill wrote:Hey, Ed, do you have a link with more info on this...and do you know if it has to be permanantly above everything else or only when filling and bleeding??

thanks

Twill
bobkoure wrote:Actually, this isn't my experience at all. I've built a number of systems using Eheim 1046, 1048, and HPPS pumps with the Innovatek agb-o-matic push-on reservoir. The fill port is the top of the reservoir, but I've never had one of these at the top of the loop - and never have had a problem, either with bleeding air (seems to be more a matter of giving the water a point to slow down as otherwise entrainment is just pulling the bubbles along) or with water coming out with the system stopped and the fill port open.
Maybe we're both misunderstanding what Ed's saying (writing)...
Actually, the primary concern is more on the issue of properly filling and bleeding out the heatercore. Sufficient flowrate will normally push the air out of blocks, but heatercores, if not mounted in such a way as to facilitate filling and bleeding, will almost always have issues getting fully bled out unless they have bleed valves in the right place. I had killer troubles bleeding out the Pro-120 in my early loops.

You can probably get away with the fillport of the loop not being at the very highest point, but I still imagine it won't be as easy as a loop designed in such a way.

I'm probably being overly cautious about it, but I had such incredible headaches from past issues that I dare not risk it with iffy loop designs.

-Ed

Twill
Posts: 34
Joined: Sun Apr 20, 2003 4:05 pm

Post by Twill » Fri Mar 11, 2005 8:11 pm

crap, now I have to re-think my loop. I have a DD fillport at the same level as my BIP II (mounted horizontally with ports facing down)

would a res fix this? how MUCH higher does it have to be.

How would I overcome the problem, any suggestions??

Twill

Edward Ng
SPCR Reviewer
Posts: 2696
Joined: Thu Dec 11, 2003 9:53 pm
Location: Scarsdale, NY
Contact:

Post by Edward Ng » Sat Mar 12, 2005 6:20 am

Twill wrote:crap, now I have to re-think my loop. I have a DD fillport at the same level as my BIP II (mounted horizontally with ports facing down)

would a res fix this? how MUCH higher does it have to be.

How would I overcome the problem, any suggestions??

Twill
Don't necessarily have to rethink it; you already have it set up, so give it a try first.

I can't tell you how much higher it must be; I don't think it's that simple. A reservoir won't automatically fix it, either.

-Ed

Twill
Posts: 34
Joined: Sun Apr 20, 2003 4:05 pm

Post by Twill » Sun Mar 13, 2005 3:04 pm

thing is I dont have it all set up yet, I still need the BIPII but dont want to shell out for a new one just yet :)

we will see how it all ends up...if it fails it fails and I re-think :)

thanks Ed

Twill

bobkoure
Posts: 191
Joined: Sun Aug 11, 2002 3:26 pm

Post by bobkoure » Sun Mar 13, 2005 5:27 pm

Edward Ng wrote: A reservoir won't automatically fix it, either.
No, but it's not going to hurt and may help.
Bubbles get carried around through entrainment, and can be released if the flow slows down. So, if it's slowing down in a radiator header tank, to the point where bubbles are being released, why not add a reservoir where it slows down even more than that. A complete fix? Probably not. Of course, given a sufficiently powerful pump, the flow in the headers does not slow down enough to release bubbles and it's a non-problem (other than it's tough getting all the bubbles out).

Post Reply