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

Props to Corsair for saying "don't waste your money buying too much of our product."

"we are going to watercool a gaming chair" That's why I'm subscribed to this channel

Linus: "do your own research" Me: "this IS my research..."

“Now, if we had sealed off every potential air leak, it’s possible we would’ve seen something closer to Corsairs simulation. But, as far as we can tell, that’s just not a true representation of the real world. I mean, nobody would do that.” I was really expecting you to follow up with “…or would they?” and then show us how you sealed up all the holes and ran the tests again.

'we're planning on watercooling a gaming chair' i thought you were watercooling a calculator already

I like the dark theme better, it doesn’t fry my eyes at 3am in the morning.

Corsair: "Dude, stop wasting your money." Linus: "B E T"

"you're gonna have to sit through this spo-" *Skips 10 seconds ahead*

Canadian air is colder and therefore has more thermal capacity. Duh.

I would have loved to see tests where both rads get "fresh" air, and also a comparison on the tripple-stack where the outside air-draw should be minimal compared to a scenario where all 3 got "fresh" air.

I just love how they didn't take the case into consideration at all. For all we know Corsair is testing in a barely/poorly ventilated case.

"Ive never done a side by side comparison" Then goes on to NOT stack radiators for a side by side comparison, and has the first rad in the loop from the MB as a exhaust and second as intake. Completely different than what he had done previously, and not at all what Corsair was saying.

"It's free" Proceeds to try to stack a single radiator on itself.

Like mentioned from others i'd like to see several tests too. 1. dual radiators with fresh air each 2. stacked radiators with the warm water in the same flow direction than the air 3. stacked radiators with the warm water counterflowing the cool air (just my guess, but 3. should work best from the stacked ones) 4. dual radiators both pulluing out of the case while other radiators push fresh air in (maybe?) Edit after further thoughts on 16.th of May to name every test that comes to my mind and compare if the arrangement is really irrelevant: 5. Dual radiators, loop in the order of CPU, radiator, GPU, radiator, pump 6. Dual radiators in the order of GPU, radiator, CPU, radiator, pump 7. Dual radiators with CPU, GPU, radiators, pump 8. Dual radiators with GPU, CPU, radiator, pump Then we should be able to see if there's any difference at all and if it matters if you place the usually cooler running device (GPU) first or not.

That sponsorship transition was so smooth I couldn’t even tell when it started

Absolutely love this video! I believe that had you mounted the 2nd radiator on a side panel as a 2nd intake radiator instead of as an exhaust radiator it would have been even more thermally efficient. Might have taken some creative cutting and mounting but I think your results would have been even better

"Nobody would do that" Says the man building a watercooled gaming chair

I think the real take away is that allowing for enough fresh air intake in a dual rad system is key to making it worth the extra cost.

Would've been nice if you tested a setup *without the radiators in series* (perhaps both outside the case, grabbing room temp air) to see HOW MUCH efficiency loss your having by pulling warm air into the secondary radiator. It seems to me that their concern is probably less that it _wont work at all_ than that it's just plain _inefficent_

7:22 the way he corrected himself tho, lolll

I'd have like to have seen Corsair's input on this, clearly they have engineers dedicated to this type of thing, and it's something they're passionate about.

Linus: does watercooling Corsair: you are doing it wrong Linus: why should I do it the right way? Corsair: be a lot cooler if you did Edit: I edited this comment so more people would see it, and *I CAN MAKE MY RISE TO THE TOP ONCE MORE....*

I think your theory about more fresh ambient air entering the case and mixing with the hot air is valid, as well as you have more volume of water now that you have a second radiator, as well as larger surface area to dissipate heat from, which add together to give you a better cooling. Thanks for the vid.

I’m liking this new type of hands-on making and testing video, more like myth busters and Ben Heck 👍🏻

Basics of heat exchange with Linus. Pro tip (literally, I'm a cooling engineer): If you want a new radiator and more cooling, put the rad in with a separate airflow path. Top and front together with a bottom and rear exhaust fan would be fine. Also, heat exchanger simulations have gotten very good now to where we see very little difference between it and the real world. The fins in the rads are tuned to optimize heat exchange with normal fans, adding to the pressure drop just makes them worse. (especially in the server example). Now, if you strapped blowymatrons to a double stack and shrouded the airflow nicely you might be into something, but one more layer of fans likely won't do it. This really boils down to Corsair talking about performance per rad and you talking about brute force cooling. They're saying you're reducing them, say to 60% performance, but you're arguing that 60% + 60% = 120% so it's worth it anyway.

ambient temperature: 21° coolant temperature: 31° * Laughs below tropic of cancer

As long the radiator is warmer the the air passing through it will have an effect. The less different in temp the less effect it will have. So best way is to have the first radiator with warmest water to start at the exhaust and last at an intake with cooler air. This should exploit the best temp difference at both places.

"You're going to have to sit through this segue." Me: This is the power of 「King Crimson」

Hold on, DARK MODE THEME ON THE INTRO!? Thanks linus for not blinding my eyes

Corsair is both right and wrong. According to heat exchangers physics, you need to "crosscool". This is, the hotter the water, the closest to the end of the flow it needs to be. In other words, the radiator the most close to the outside of the case needs to receive directly from the components (hottest water) and the one most inside (closer to the most cooling active components, like fans, needs to have the coolest water, after being being sequencially and in series piped from the hottest to lowest. Stacking rads is the most efficient way to do it. If you do know how to do it. Rads need to be in series. Hottest with less flow closer to the end of the case. Output of coolest water on the most inside rad. Source: nuclear reactor heat exchangers. Edit: Clarification and correction of incoherences. Also, thanks you guys. Lost the LTT heart, sad. It was an honor to have it, though, even if it was only for some minutes. Thanks LTT.

Missed an important point: The order the water & air flows through the radiators. For best transfer you also need to ensure the hot water from the CPU/GPU goes through the 'exhaust' radiator first, then to the 'inlet' radiator. That way the hot air is passing through the hottest radiator & the precooled water gets the cooler air from the inlet, AND means that the heat from the inlet cannot increase the temp in the outlet radiator. Order is important!

As an AC tech I had an odd thought you mentioned those rads are in series I'm thinking it might make a difference which one is first in line getting the cold water

"you're going to sit through this se-" *presses right arrow three times*

If you stack radiators directly on each other the flow direction is important. The one that is closer to the outside should get the hot water first(if the air is generally flowing out (for example top mounted radiators blowing hot air at the top out.) . The water is after that only warm and now gets to the second radiator more inside, that has more cold air. Fish use that same principle to maximize oxygenation in their gills. In German it is called the "Gegenstromprinzip"

It makes me wonder if it was the second radiator or rather the extra air flow from the additional fans. Have you tried any tests where you left the extra fans but not the extra radiator?

Well done on the testing and proof! Always love the content!

The real question is, is 2 units in paralell pulling in air from their own side (say top and front)better than the config shown in the video

The last time I heard about XSplit was when streamers complained about it for being so expensive before switching to OBS. Good times :P

I would be interested to know the coolant temps entering and leaving each radiator. I would be willing to bet that the exit coolant temp of the first radiator is still higher than the exit air temp. on the stacked setup I would bet the best config would be to have the air exit side be the coolant hotside and the air entrance be the coolant cold side.

Hey Linus, I reckon it would be great if you could write a paper on this with a research question that falls along the line of effectiveness versus efficiency. 😊

Someone is losing sponsorship from a certain computer accessories company for the next few months.

Linus: “Do your own benchmarking” Me: “okay, let me order what I need“ My Bank Account: “Bruh, not even close”

When 400 watts usage was impressive lmao

Please dont rinse me here, but i recently built my pc. Im using an NZXT H210i with two internal rear fans (one top, one back) both set to exhaust and a front 240mm AIO rad with fans drawing air into the case over the rad. Have i got the front fans the wrong way around? Should i be blowing air from inside onto the rad so it immediately exhausts it? The front of the case stays the most dust free though as far as ive found

The thing here is that the temperature gradient is not as high for the second radiator as for the first. therefore the second rad won't be as efficient as the first, but that does not mean that it has no effect. and stacked radiators do not slow down air enough to "hold back the heat", especially not if you make a rad-fan-rad-fan-... sandwich. if you use a fan-rad-rad-rad-... setup there will be a point where more rads will negatively impact thermals, but that would need tens of radiators so you'll never reach that even in a server scenario. bottom line: the effect corsair points out is a thing, but it's not relevant for real-life usecases.

The only negative effects I can think of too many rads is, diminishing returns. 🤷🏻‍♂️

Did you calculate the increased specific heat capacity of the water in the additional loop to see how it offsets the time to temperature?

Would be really awesome if you measured the air temp or coolant temp after the radiators with a thermocouple, especially for like a stacked radiator setup to test how much cooling the 2nd and 3rd rads added

Corsair: Buy less of our equipment, might help Linus: Nope, buy everything they have, will help Weird marketing strats.

When the beard goes, I go.

HEY LINUS! pro question!!! If you have 2 rads cooling the loop, is it better to put the output radiator between the CPU and GPU then let the cooler rad cool the water going to the water pump? or is it just better to put the radiators after echother in the loop? Seems like there is a lot of ways to still use at least 2 radiators. But if you put one between you till be able to deparate the heat imput and cool more efficient? Or am i wrong?

Guys, I need to mount my radiator to the front, because of RAM clearance. HHowdo I mount the two fans on the radiator? Should they pull or push? And: On which side of the radiator should the fans be? Is it "Case -> (inside) -> radiator -> fans -> ..... " or "Case -> (inside) -> fans -> radiator -> ...." ? I have two other fans that are pushing air out of the case, one at the rear top and one at the top rear.

Corsairs' rad calculator is basically right but likely lacked the required data/specs of the exact cooling system units you used, so the substituted it with "generic data", Their heat exchange calculator is likely based the same type of calculations (thermo-dynamics & fluid dynamics) used for building HVAC Systems and Vehicle radiators/intercoolers when perusing maximum cost to efficiency of any heat exchangers. A. Their feed back on your 2 systems were likely based on: 1. The HackPro (1 front intake rad & 1 rear exhaust rad), where the first rad pulls cool external air to cool the first loop. This results in the air leaving the first loop and, being hotter than ambient.. increasing the temps within the case. The now already heated air inside the case, was then used to "cool" the second rad, which then finally expelled the even hotter air out the back. So though it may not necessarily be pointless (depending on the spec/parameters of the loops themselves, additional exhaust ports, etc), it is definitely not the most efficient cost to performance setup. Simplified solutions - a. Add an intake fan to the top of the case & flip the front rad to blow outward or b. Add an exhaust fan to the top of the case & flip the rear rad to suck in 2. The Custom Minecraft server (Stacked rad. build) It follows the same principle above increasing the temp of the air leaving every rad then entering into another. It's just a matter of inefficiency. Simplified solution - Take the exact series rad configuration and turn them parallelly with a radiant barrier between them and you'll see a greater value. I'm unable to say just how much gains you'll see without knowing the full loop's specs.. but yeah, you get the idea. B. And finally, the reason you got a better reading with your test system at the end of this vid, is actually kind of simple and obvious.. You actually used 2 feed separated, perpendicular intake systems that both pulled cool ambient air thru them, with case exhausting the hot air thru its rear. So it pulled from 2 cool sources and shared a single exhaust point (which is near always the best solution for this type of scenario). If you used an exhaust fan at the back, it would increased your efficiency even higher, as it was only operating on compressive force (a high positive pressure) within the case to discharged the hot air, rather than a "guided controlled flow". Again, there are actually a calculations for sizing the fans to give a desired low positive pressure, or even a negative pressure and determining what is the best to maximize efficiency for each scenario. Think of it like this.. when increasing voltage for overclocking, there comes a time when the increase of voltage consumption vs the gains seen aren't efficient anymore.. even if minor gains are there.. well this is basically the same thing. Heck, even these radiator systems could be made more efficient with the introduction of thermostat units (like nearly every other rad system).. but I'm sure on this small scale they are left out because of the cost/savings just aren't worth it and they are already able to reach the required thermal control.

"It's Easy, Cheap, and Free!" if you have a warehouse of liquid cooling parts laying around. Linus! Let me play in your warehouse!

7:05 Is it less that its a bad thing and more that its "far" less efficient as to having both radiators as say, intakes and exhausting out the back.

Question, have you had a radiator leak? Mine did last year and since then I've moved my radiator to the front of my pc instead of the top of my pc. Granted my new case has that ability whereas my original case did not. Managed to get the parts all reworking even after leaks and shortages, minus my ram I think thats finally kicked it last week and needs replacing. That or the MOBO

LTT: has been water cooling wrong for years Manufacturers: *_that's because you're doing it wrong_*

Linus: You're gonna have to sit through this segue. You underestimate my skip button.

"you are going to have to sit through this se-" me - KING CRIMSON *intro starts*

To understand the basics of the engineering behind these arguments, a good resource is the datasheets at Koolance for cold plates, heat exchangers, and pumps. Without having to do the heavy lifting calculations, you can observe from the various data plots that total fan CFM is a major figure of merit behind the maximum TDP. Flow rate and radiator size (thickness, number of fins) all appear to be secondary contributors. Radiator area contributes in the sense that it potentially adds more fan CFM.

“to find out what they said, you’re gonna have to sit through this seg-“ *skips ahead 15 seconds*

Still waiting for water-cooled linus

Short version: you can't cool something with warm air! Makes a shitload of sense to me.

I would be curious to know the actual rad temps. Stacking rads seems like you would heat the second rad with the hot air from the first. But I guess if the air moves fast enough that’s not an issue.

corsair: "using rads behind each other in a airflow starved case is bad" corsair: "using rads stacked on top of each other is bad" linus putting another rad in a case with a completely open back: "look how much it improves cooling!"

"You're gonna have to sit thr-" *Hits right arrow 3 times* :)

What was your water flow through the radiators? Did you run the water through the top radiator first and then the front one? or the other way around?

Would love to see if the order of the rads in the loop would make a difference as well. If the exhaust rad was hit first by the warm coolant it would increase the temp difference there, but on the other side lower it for the second rad. I wonder if these two balance out or if one of the two has a bigger effect than the other...

I think maybe Corsair was trying to make a diffrent point here .... cooling would be better having the radiators side to side than face to face. Also, what if you put both radiators taking fresh air from outside and the vents just taking the hot air out? PS:Are you conecting the radiators in parallel so the IN of the second one is connected to the OUT of the first one? Is the hot water going into the 1st radiator first or into the 3rd radiator first? I love you guys but sorry, your experiment is not very scientific. Not a really good empirical test I beleive

Linus should hit up Engineering explained and look up the Honda Civic modding community to do some research into the physics of stacking radiators

You should do a countercurrent exchange to improve efficiency. If you put the radiators in series and have the hotter one as the exhaust, that should work.

Corsair: "stacked rads are barely effective." Linus: explains why that makes sense and decides to test it. Also Linus: doesn't test stacked rads. Even alsoer Linus: Explains that the stacked rad setup only gained 5 degrees compared to the 20 the serial rad system dropped. Even more alsoer Linus: "iTs tHe SaMe ThInG."

I'd like to see Corsair do an experiment and made a video call: $h!t that Linus said

"just like in the simulations" -Cosair

Yeah, I just went through this with the case being in the way of the rad and fans. I put the fans on the outside of the case, Rad inside the case. Didn't work well. I cut the top of the case grille out, the remounted the Rad on the outside of the case. Then out the fans on top, drawing air out of the case, got +3c to cooling, even more significant at higher Temps. Its all about airflow over the rad. More Flow, more cool. Also, cooler air = more cooling and heat Transfer/exchange.

Did you keep the same amount of fans for both instances, or for the double radiator did you install more fans?

I’ve been water cooling wrong for YEARS.....says the guy who water cooled his whole room

The big thing is, from what i can see in the emails from corsair is that they are critiquing the stacked radiators with fans not a case with multiple radiators in different places.

ah yes the easy, free, fun benchmarking marking i can do with all the spare h100i's i have lying around

LINUS HELP I’m getting a i5 9600k and I was wondering what aircooler you would recommend ? I have a med sized case with not that much space

"It is not efficient to stack radiators" i think this is the correct statement

I didn't get it :( Corsair was talking about stacking radiators one onto another so that the heat from the first radiator goes through the second radiator (not cooling it as result). So you would need to compare two stacked radiators vs two "chained" radiators to see how much stacking affects cooling efficiency, no?

"You're going to have to sit through this..." *four right arrow key clicks later*

Will this work if I have a positive pressure system and 4 radiators, 2 on the in takes and 2 on the out takes?

“I’ve been water cooling incorrectly the whole time” LIEnus confirmed

There’s one other possibility: the assumption that radiators are near 100% efficient and effective is wrong.

I have lots to learn when it comes to computers and I hope not to insult, I really like your videos. I am left with a bigger question, would 1 radiator of equal size of the 2 used in second test run cooler? It didn't appear to me as a fair test. I am going to be attempting to build my first water loop soon and that answer would help me decide which case to go with.

@Linus, would be interesting to see if there was a difference in temps if on the stacked rads you reversed the progression through the radiator. i.e. 3rd radiator being the exhaust side as the first radiator in series (hot coolant in) then the sandwiched rad, then the intake rad. In theory this would give you the greatest amount of potential drop across the assembly, whether or not that translates into real world differences in temps.... *shrug*

Actually Corsair is right. WHEN YOU stack radiators and DON'T CONFIGURE them in COUNTERFLOW, meaning when you hot water enters the coolest part or your radiator (the first radiator). In this case ur water is cooled down and the air is warmed up. So far so good. But at the next radiator ur cooled water is forced to deal with the already warmed up air. This Effekt faster gets to a point where your cooled water can no further be cooled because of your warmed up air. You will better prevent this when you feed your hot water into the hot end,meaning the last radiator. Then your hot water is cooled with the already warm air from the other radiators. Then the cooled water is further cooled by less hot air. And at the first radiator your chilled water is cooled by the most possible cool air, ambient temperature. Using this called counter current model provides the best temperature differences between water and air regarding the complete cooling system. Because ur coolest point of water is at the coolest point of your air. That's why counter flow models are widely spread in the chemical industry, where cooling efficiency is lots of money and a security issue.

Corsair: in this test cases another radiator is useless Linus: let me prove you're wrong in a different scenario

at 1:30 you have shown an image of the case that we can call as a loss as mechanical engineers. you can say that 25c air is passing the first radiator at a temperature of 30c, entering second radiator at 30c existing at 32, entering the 3rd radiator at 32 exiting at 33c. we can expect water temp as following : at rad 1.... 40c in, 35c out, at rad 2... 35c in 33c out, then at the third water will reduce almost out at 32. so the Q the heat you sucked out by the first rad have been good and the watt you used to do it is small when you calculate the Q of rad 2 and rad 3 compared to the same watt which is fixed to circulate water across a single rad. the best thing is to create air ducts by 3d printing as example giving fresh air at 25c to all rads and same duct to suck each rad hot air out of the case on another direction away off the intake. if you are interested, make someone comment back to me, i can give you some drawing or 3d drawings for printing based on your rad dimensions

Would having both radiators as intake and using all leftover fan space in the case as exhaust make for a better cooling, over running one as intake and one as exhaust?

ambient temperature: 21° Me in Chennai,India: "Oh my! These guys must run some pretty ballin industrial grade AC at full blast for these tests!"

Process engineer here, I often do chemical simulations involving heat transfer. I think I kind of get Corsair's point, even if the wording is not correcto IMO: if you add more dissipation surface, the temps will always be lower. However, for a fixed exchange surface if the temperature profile is flatter (or DTLM as is known in heat exchanging), the total exchanged heat will be lower, leading to less bang for your buck. You can even get a reduction in heat exchange if your temperature profile difference is too high depending on the exchanger thermal approach (the effect is less wanky and counterintuitive if there is no fluid phase change, like here). Also there will be some effects on the boundary layer size impacting the boundary layer conductive and convective coefficients, but I'm not going to touch that with a ten foot pole as it's way outside my field of expertise.

Does the water flow parallel or in series through the radiators and if it it series which radiator is first?

Yo thanks for the nightmode intro That was nice at 11 pm

Now to wait 8 months for Linus’ “I was wrong” video.

finally, they put their intro in dark theme, everyone knows its superior than light theme

I have E850 power supply, and I'm just wondering are temperature above 40C safe for it. Because it ranges between 30C-45C?

did you try to put one radiator on the input of the graphics card an the other on the output for cooling the water that is warm because the dissipation of the graphics card and cool it another time to put in to the cpu ?? I think that the orientation of the fans can be with all the fans of the radiators pushing the air out of the case, and 2 or more fans that u have pushing air into the case, or maybe can be reverse.