Dual 120, layered?

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1
Evening folks, has anyone ever tried to use two 120mm radiators layered, with a fan in between, so effectively get 240mm worth of radiator?
What I'm visioning is using two crossflow radiators, and put them in series (input into the first rad from res, output from first into input of the second, output of the second into CPU) to basically make a really thick rad, with a fan in the middle.

the purpose of this is to use in an ITX case, where there's only a single 120mm mount, but there's more than 15cm of depth.

Anyone done this before? or will i be one of the first mad enough to do so?
 

EndEffeKt

SFF Lingo Aficionado
Mar 23, 2019
106
34
Never seen it but sounds interesting. You will need to tinker some kind of shroud to direct airflow.
 

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1
The plan is to do something similar to push pull, but instead of "fan-rad-fan" it'll be "fan-rad-fan-rad", since i figure one fan can't deliver enough airflow. The fan between the rads will somehow mount to both rads, so that should do it for directing the airflow. On paper it sounds feasible, though until i get the rads and fans home, i can't test it
 

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1
You could use ultra thick rad with 38 mm or even two 38 mm fans.
From what i can gather online, having more radiator is better than having a thick one, hence why i'd try to layer two 120mm with corresponding static pressure fans.
I don't know what you mean by 38mm though, is it the thickness of the fans?
 

paulesko

Master of Cramming
Jul 31, 2019
415
322

read this, it´s interesting. Maybe its more efficient to use I super high fpi rad like the koolance with two fans than the other way round. Using two rads and one fan seems like the worst use of space.
 

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1

read this, it´s interesting. Maybe its more efficient to use I super high fpi rad like the koolance with two fans than the other way round. Using two rads and one fan seems like the worst use of space.
Interesting, thanks for finding that. i couldn't find anything myself
My plan would be having to rads and two fans, like the 6th version he's showing off, so i was hoping to have to slim 120mm, with two static pressure fans, to increase the capacity of my loop, since I'm hopeful that i can cool CPU & GPU

 

paulesko

Master of Cramming
Jul 31, 2019
415
322
I think that if you have the place to put that, maybe it´s better (as shown in the article) to use a normal rad with two fans. ANd I guess that if you find I very high fpi rad with two fans would work great, but it´s just my opininon

look at this koolance 30 fpi 54mm thick rad. If you put two high pressure fans in push pull, this has to be a killer setup

anyhow the thing would be between a single rad with very high fpi and quite thick (30 fpi koolance) in push pull, or two very low fpi rads with two fans... I think black ice have 9fpi rads. This is an experiment I would like to make, but right now I´m into too many things at the same time.
 

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1
I think that if you have the place to put that, maybe it´s better (as shown in the article) to use a normal rad with two fans. ANd I guess that if you find I very high fpi rad with two fans would work great, but it´s just my opininon

look at this koolance 30 fpi 54mm thick rad. If you put two high pressure fans in push pull, this has to be a killer setup

anyhow the thing would be between a single rad with very high fpi and quite thick (30 fpi koolance) in push pull, or two very low fpi rads with two fans... I think black ice have 9fpi rads. This is an experiment I would like to make, but right now I´m into too many things at the same time.
I think I'm gonna experient around a bit when my rads come, to see if it's beneficial. Might pick up a few more with different fpi, to see what benefit there is.
Really regret throwing my old testbench out now??
 

giraffesinmybalcony

SFF Lingo Aficionado
Dec 15, 2018
95
88
im pretty surprised no one has mentioned this so far, but using two rads sandwiched together is a rather pointless idea. if you need further explanation you can find it online, im too lazy to type it all out on my phone
 

AP2

Airflow Optimizer
Feb 1, 2017
291
325
I did this recently and I was able to cool a 8700k and 1080ti Mini. Temps were pretty good running the Noctua fan at about 1000rpm.

I went this route to get a 240 rad size in a Dan A4/Thor case.

 
Last edited:
  • Like
Reactions: Phuncz

paulesko

Master of Cramming
Jul 31, 2019
415
322
yes, but those are 9fpi rads, aren´t they? In my opinion probably the only radiators you can use in this kind of setup
 

RoedCrimson

Efficiency Noob
Original poster
Sep 12, 2019
7
1
thin
im pretty surprised no one has mentioned this so far, but using two rads sandwiched together is a rather pointless idea. if you need further explanation you can find it online, im too lazy to type it all out on my phone
thing is, i can't find anything online, hence why i've asked here
 

giraffesinmybalcony

SFF Lingo Aficionado
Dec 15, 2018
95
88
thin

thing is, i can't find anything online, hence why i've asked here
here you go



edit:: the diagram you posted above is literally from a forum post testing the radiator setup with conclusion that stacked radiators almost always perform worse than a solo radiator. just intriguing that you said you cant find anything but i hope that helps with your question.
 

TristanDuboisOLG

Average Stuffer
May 10, 2018
81
21
I've done lots and lots of searching on this subject when I was discussing alternate layouts and features for the S700 case with Salvo.

Basically, you have to break this down into simple thermodynamics. If you're cooling something, you're stuck at a cooling rate determined by effectiveness of the design, (fin stack (air) or fin density (water)), and the max thermal conductivity of the medium (air or water). The reason there isn't that much difference between air coolers and water coolers is because the heat from the CPU is eventually going to a copper stack and being sent from the stack to the air exhausting the case.

There have been several interesting tests done on this on Martinsliquidlab.org back in 2012. I was fortunately able to read them before they went offline, but there are a couple of other less in-depth threads about it here.

While your idea is interesting, it unfortunately won't work. Here's why:

When you sandwich two radiators together there are a few things that matter, loop order, fan placement/configuration, and radiator size.

When you try to cool your system, cool air will flow over radiator 1 and absorb as much heat as it can from the fans. When it passes through radiator 2, the air will have lost a lot of it's thermal potential when it passed through radiator 1. Thus, not really cooling. In fact, in a lot of the tests they found that 2 radiators had negligible improvements over having only 1. (Partially because if you think about it, radiator 1 gets all the cooling, radiator 2 is being heat-soaked).

People thought that they could possibly increase performance by moving more air in using a (Fan | radiator | Fan | radiator | Fan ) setup but it doesn't seem to move more air, it just seems to make the flow choppy.

I would guess that the only way you could do this and see any kind of performance boost is by using 2 extremely thin radiators (10mm) and stacking them. This way you could separate and cool 2 loops, (take up roughly the same fins stack area as you would normally), and in the cooling process you wouldn't need to worry about heat contamination from GPU to CPU.

It would have to be a very unique use case for it to be a plausible solution. Sorry :(
 

temu

Case Bender
Oct 3, 2019
2
0
It's a decent idea. Two half thickness rads in series and stacked as you propose give better heat transfer than a single full thickness rad. With the cooler rad at the front receiving the cooler air and the hotter rad stacked behind receiving the hotter air, you'll reduce delta T at the front but increase it at the back. Because the temperature profile of the air flowing through the rads is not linear (it asymptotes toward the fin temperature) the increase at the back will more than compensate for the reduction at the front, and you'll get a net increase in heat transfer.

It's similar to the way counter-flow gives better heat transfer than cross-flow in normal heat exchangers (and thermo 101 for a lot of people, who are rolling their eyes at me).

However, the water temperatures in most cooling loops don't drop that much in the rads (a consequence of high flows used to get heat out of the cpu and gpu blocks). You might get a temperature difference of 1 or 2 degrees, at best, between the rads. With such a small temperature difference between the two rads, the gains over using a single full thickness rad will be pretty small.

You can see this in the results shown in that xtremesystems link. Putting the hot rad in the front (#7) should have the opposite effect of putting the hot rad at the back (#6), so the performance of a single rad of the same total thickness should lie between #6 and #7. But #6 and #7 are indistinguishable at 800rpm and barely different (0.3°) at 1200rpm. So the effect of using two thin rads instead of one thick rad, for that configuration, is weak.

And the rads in #6 and #7 are obviously starved of air pressure (compare them to #2!), but this is probably the more important issue - balancing rad air resistance and the fans pushing against it. I think this might give a better reason to use stacked rads.

You've 15cm. That is a lot thicker than what most rad/fan combinations are optimized for, so filling it with a thick rad and two fans in push-pull will probably not be that close optimal - you probably won't have pressure to create the airflow to get much heat transfer out of the back half of the rad, or you'll need screaming industrial fans, or you'll need low fin count rads that don't give enough heat transfer. Filling it with two thinner rads in a fan-rad-fan-rad-fan configuration should give you more pressure that could give you enough extra airflow to compensate, somewhat, for the reduced cross section.

Or it might not. I do think it's worth trying, though.

Edited to (try) reduce cruft.
 
Last edited:

temu

Case Bender
Oct 3, 2019
2
0
Oops, that wasn't so helpful.

If you want 240 thermals out of a 120 area, you can try the rads you've got. If they give you double the thickness of the 240 you'll want a lot more flow (ideally double) and need to increase pressure even more (because both the thickness and the flow are increased). If a second rad doesn't make a difference, you can try increasing pressure with more/deeper fans or rpm - whatever it takes to overcome the air resistance and get cool air flowing to the back rad. Adding even more rad thickness will reduce the required flow, but might increase the required pressure. Reducing rad thickness, the opposite. You might find a combination that meets your requirements.

There may also be thick rad and fan combos designed for your purpose. But I doubt they're optimal. It's a long way from the common rad fan combos.