I’ve been water cooling wrong for YEARS - $H!T Manufacturers Say

[ColinLTT]

Apparently stacking radiators is bad -- According to the brains over at Corsair, so we built a water cooled gaming rig to test their theory.

 

 

 

 

Buy Intel 9900KS
On Amazon (PAID LINK): https://geni.us/RyZyZQ5

 

Buy Corsair Vengeance RGB RAM
On Amazon (PAID LINK): https://geni.us/uxWIG
On Newegg (PAID LINK): https://geni.us/3HfM0Ii

 

Buy EK Reservoirs
On Amazon (PAID LINK): https://geni.us/YubShVJ
On Newegg (PAID LINK): https://geni.us/kPLy

 

Buy EK Coolstream SE 240mm
On Amazon (PAID LINK): https://geni.us/bZQBl
On Newegg (PAID LINK): https://geni.us/VUjAIu

 

Buy EK Vector RTX 2080Ti
On Amazon (PAID LINK): https://geni.us/aeSlS
On Newegg (PAID LINK): https://geni.us/axcLVv2

[Tristerin]

I might watch this video, Ive thought about stacking a series of 240mm rads I have but the only thing that was stopping me is how horrid the zip ties look to get them to sandwich together (didnt dig into other ways to get them stacked with fans inbetween yet)

[samcool55]

Not a scientist but my theory is that corsair's theory is only true in 1 scenario.

It can only ruin temps for one loop. If you have 2 loops, 1 very hot loop (let's call it loop A) and 1 very cold one (loop B). And a specific rad config (a very stupid one).

If you place the rads for the loop as follows: fan -> rad loop A -> fan -> rad loop B -> fan.

My theory is that once the air leaves rad A, it's hotter than the water temp of rad B which causes loop B to heat up instead of cooling down.

But eventually you will reach a thermal point in loop B that's likely around the same temp as loop A and from that point it will no longer keep heating up.

 

It's basically heating a loop with the heat from another loop at this point.

[Davy12]

I joined the especially for this forum since I really like physics but I can really follow corsairs theory here. But in reality their theory only makes sense when the radiators are mounted really close together with bad airflow I think. There has to be a balance between the amount of airflow which decreases with an increasing amount of radiators (if you do not increase the amount of fans) and the amount of radiators. I would really like to see how the system would perform without an increasing amount of fans and just putting two radiators behind each other. I think the results will be less clear, sure the temperatures would rise slower, but at thermal eq. It should be just the same if the radiators are airtightly connected. 

So to be short the test of Linus was doomed to fail (establish his theory) since you did not only increase the amount of radiators but you increased the amount of airflow.

[ColinLTT]

2 minutes ago, Davy12 said:

I joined the especially for this forum since I really like physics but I can really follow corsairs theory here. But in reality their theory only makes sense when the radiators are mounted really close together with bad airflow I think. There has to be a balance between the amount of airflow which decreases with an increasing amount of radiators (if you do not increase the amount of fans) and the amount of radiators. I would really like to see how the system would perform without an increasing amount of fans and just putting two radiators behind each other. I think the results will be less clear, sure the temperatures would rise slower, but at thermal eq. It should be just the same if the radiators are airtightly connected. 

So to be short the test of Linus was doomed to fail (establish his theory) since you did not only increase the amount of radiators but you increased the amount of airflow.

The number of fans did not change during testing - only their location. The two fans you see top mounted in the dual rad test were internally mounted for the single rad test. 

 

[Davy12]

1 minute ago, ColinLTT said:

The number of fans did not change during testing - only their location. The two fans you see top mounted in the dual rad test were internally mounted for the single rad test. 

 

Ah so very interesting. But it could disperse the heat two times as fast since it has two radiators which we can thus, as proven by linus, view as nearly independent cooling systems. But I would still be very interesting to see the same set up with two radiators mounted behind each other.

[Doommius]

So, i just watched the video and i have 2 comments, i would say water flow has a lot to say here as well, if you the loop goes:

CPU -> GPU -> front rad(intake) -> top rad(exhaust) -> pump -> 

vs

CPU -> GPU -> top rad(exhaust) -> front rad(intake) -> pump ->

 

as the last will will be able to cool batter as it has a higher delta T for both rads.

 

Just my two cents. would like to see something like this as well for the Minecraft server build with flow back to front rad and front to back rad.

[RandomPhysicist]

4 minutes ago, Davy12 said:

Ah so very interesting. But it could disperse the heat two times as fast since it has two radiators which we can thus, as proven by linus, view as nearly independent cooling systems. But I would still be very interesting to see the same set up with two radiators mounted behind each other.

I agree i also just joined to warn linus of this, as can be seen in that glimpse email from corsair, they were commenting about the "stacked" radiators as in fan and rads next to each other ziptied with almost "no airflow in between". In the video you added another rad not stacked up againts it meaning yes by your theory air from the outside influenced the internal temprature making it work however, in the situation of the mc server where theyre stacked and have no breathing room it will be entirely different 

[riklaunim]

I have two 240 radiators as Alphacool kits. One for Vega 64 and the other for 1920X. I wanted to put the Vega radiator on top of the case as it would be hottest more often while gaming but the tubes weren't that long, so I had to put it at the front panel and CPU at the top.

At max synthetic synthetic load the whole PC got kind of hot (like 60C). What I could also do is combine the loops due to Alphacool connectors (vega -> rad -> CPU -> rad ->). That did lower the temps on the GPU while CPU was bit warmer due to that.:

 

 

Yet the fans had to work on mid-high RPM. Now I have the rig on Streacom BC1 and the system is pretty much noiseless as both rads get ambient air and use lower RPM.

[Vortal]

I don't think your bench test is valid to demonstrate Corsair's point

1)most the heat generated in the box (CPU and GC) is going in the water, not in the case. the only heat you are picking up is maybe the RAM, VRM, and PSU... you can check that mesuring a delta temperature between inside and outside the box.

2) from what is shown on the video, 2 radiators stacked one behind the other is what is bugging Corsair. and rightfully so. puting them in series makes rad #2 see all the hot air of rad #1... ending up having 2 rads performing almost the same a 1, (but paying for 2)... what would REALY be efficient, is having them set in parralel. one next to the other both seing fresh air. thats were having 2 rads becomes worth it.... 

[Dravinian]

I would like to see a better video of this.

 

This one seemed, 1, extremely short, 2, not very scientific, and 3, limited to a scenario which was entirely different to the one that corsair highlited as being an issue.

I would like to see a recreation of the actual case builds that corsair highlighted as being a problem and then a proper gamult of testing to see whether there is actually any difference, with perhaps some variations thrown in.

 

My summation is: good idea for a video, given current situation, probably wasn't in a position to fulfil the potential.

[Davy12]

3 minutes ago, Doommius said:

So, i just watched the video and i have 2 comments, i would say water flow has a lot to say here as well, if you the loop goes:

CPU -> GPU -> front rad(intake) -> top rad(exhaust) -> pump -> 

vs

CPU -> GPU -> top rad(exhaust) -> front rad(intake) -> pump ->

 

as the last will will be able to cool batter as it has a higher delta T for both rads.

 

Just my two cents. would like to see something like this as well for the Minecraft server build with flow back to front rad and front to back rad.

Once the system gets to thermal eq. Which we saw was pretty fast under high load the loop itself should not matter anymore I think.

[Davy12]

1 minute ago, Dravinian said:

I would like to see a better video of this.

 

This one seemed, 1, extremely short, 2, not very scientific, and 3, limited to a scenario which was entirely different to the one that corsair highlited as being an issue.

I would like to see a recreation of the actual case builds that corsair highlighted as being a problem and then a proper gamult of testing to see whether there is actually any difference, with perhaps some variations thrown in.

 

My summation is: good idea for a video, given current situation, probably wasn't in a position to fully the potential.

I agree I want to see a full scientific study and the results better have some fancy error bars too.

[masamoto]

I really think this whole thing has more to do with. Corsair Simulating with the "Perfect scenario". Where the heat exchange on their intake rad  that is first in the loop is enough to heat up the air that a second radiator that acts as an exhaust rad can not dissipate any more heat. It's just the wrong setup. I would like to know how your chassis temperatures were. And I am not a water cooling guy but hear me out. Would having the exhaust fan as the first radiator in the loop. Using the warmer but still cool enough air in the chassis the get some heat away from the loop. And having the intake fan be the second in the loop. Having to deal with lower temps and cooler air not be the optimal solution? I think about that with the radiator stacked server as well.

[haxfar]

4 minutes ago, Doommius said:

So, i just watched the video and i have 2 comments, i would say water flow has a lot to say here as well, if you the loop goes:

CPU -> GPU -> front rad(intake) -> top rad(exhaust) -> pump -> 

vs

CPU -> GPU -> top rad(exhaust) -> front rad(intake) -> pump ->

JZ2cents did a video on it and found negligible difference

 

[masamoto]

Just now, haxfar said:

JZ2cents did a video on it and found negligible difference

 

He is not talking about the general order in the system. More about who gets to dump the heat first in a closed chassie. Jayz video with the open workbench is something different.

[Doommius]

Just now, masamoto said:

He is not talking about the general order in the system. More about who gets to dump the heat first in a closed chassie. Jayz video with the open workbench is something different.

 

3 minutes ago, haxfar said:

JZ2cents did a video on it and found negligible difference

 

 

8 minutes ago, Davy12 said:

Once the system gets to thermal eq. Which we saw was pretty fast under high load the loop itself should not matter anymore I think.

Yeah, I'd assume doing something like this in a close cabinet,

[haxfar]

+1 on what exactly is corsairs mesh, physics setup and results.

 

Reporting seemingly no difference between side panel on or off is a bit suspect and imo warrants further looking in to.

 

I would like to know if there's a significant difference between running a 2* 240 and 360 +120.

The 360 should give a positive pressure and only leaving heated air for the exhaust, whereas the 2* 240mm (and possible location) might have the exhaust able to pull in non-heated air.

[Davy12]

1 minute ago, Doommius said:

 

 

Yeah, I'd assume doing something like this in a close cabinet,

I wonder why would you have some fans sucking hot air in? I would mount all those fans going through a rad going out. Get some fans on the back sucking fresh air in.

[Doommius]

Just now, Davy12 said:

I wonder why would you have some fans sucking hot air in? I would mount all those fans going through a rad going out. Get some fans on the back sucking fresh air in.

Dust, most cases have filters front and bottom. 

[Davy12]

Just now, Doommius said:

Dust, most cases have filters front and bottom. 

Ah, shows I'm a pc noob, but why not just make the bottom one's blowing in?

[Nystemy]

Modeling thermal behavior is frankly far from trivial.
Adding a second radiator will generally increase surface area and offer more potential for the air to snatch heat from the water.

The second radiator will also add more thermal mass, but that is another can of worms since it isn't useful for all workloads.

 

Though, I can't agree with Corsair's statement that the air will be nearly the same as the water in the loop, since then one must have a fairly efficient radiator design. (Either thick radiators or low air speed, or a combination) How large the temperature difference is in practice can be measured so it can be worth while taking a look at. (Airspeed must though be taken in consideration for it to be a proper study though)

Also, placing radiators so that the warmest radiators gets to see the warmest air is usually a good idea to implement. (counter flow heat exchangers are a rather useful design to follow.)

In the end, as long as the radiators aren't restricting air flow (ie have fans), then it shouldn't be a major issue to toss in more of them.

[Doommius]

Just now, Davy12 said:

Ah, shows I'm a pc noob, but why not just make the bottom one's blowing in?

In this cases they were testing normally you would probably do this if you aren't in a super dusty environment, personally i'm sucking in front and button, exhaust by the cpu in the back and side panel taped over with duck tape as to not suck dust in.

[sacharja]

Does anyone know what case was used there?

 

Am using a *fanless* PC and looking for an open / mesh case that supports natural convection, has enough space for my big passive cooling blocks and some 5,25" bays to update usb front connectors in the future. Seems that the case in this video would suit.

[GDRRiley]

Corsairs model seems to be over estimating how efficient the rads are. I don't think any rad would take say 40C water with 20C air and turn it into 39C air and lower the water temp. There also maybe a point that a single bigger (360 or 420 or 480) is better than 2 240 but that would mean a lot more testing which I'd love to see.