I have a background in heat transfer, just to piggyback that the amount of heat flow gets dissipated into ambient via a rad plateaus really quickly against the fin length (rad thickness in this case), it's actually a tanh curve if I'm not mistaken. In other words, once passing a certain height (which is determined by many factors itself), you are not going to notice any meaningful gains, which is demonstrated in this video. In convection heat exchanger, fin density and length of the rad matters much more. Honestly high pressure head fans at low RPM with low resistance rads works wonder.
This stuff is so fascinating. I’ve been playing around with a custom loop and two 240 rads for the last year and I’m STILL getting surprised at how certain configurations manage heat, air flow etc. Not to mention the case itself impacts the end result too ❤❤ it’s all one big beautiful complicated, yet understandable, mess 😂
yes, all else being equal, but generally, you have variable core amounts, variable restriction, double stack fins vs single. I can have seen thinner denser fin stacks lose to thicker less dense fin stacks in bennchmarks at various speeds.
most valuable video about picking a right radiator for your CL .... this was pleasure even to scroll back and listen again and again... dude more of thise dude MORE !
I've seen this data replicated over the past 2 (or so) years and found it still holds up. In my SFF builds, cooling comes down to the fans. The dirt cheap radiator has always worked better in combination with the Noctua fans. And yes, the pumps are a factor. Indeed poor pump leads to poor results and elevated temps. AIOs are a good example (exit costume loops) since some rpms are "fixed". Love the info and please keep it going. And, of course, thanks for the content!
Thank you so much for the thorough stuff, concise explanation on findings, accurately representing graphs that look sweet, and the room level speaking volume. My ears and eyes appreciated this video.
Well presented very interesting topic. I really enjoyed to watch it. Though I can't totally agree with conclusion. The thing is that in order to answer the question of rad thickness affect on cooling in a more scientific way, some conditions should be met. And the most important is that the rads should be all from the same series of the same manufacturer. This is because the difference between various models can be huge. For example the 30mm HWLabbs Nemesis GTS 280 will slightly outperform 45mm EK CoolStream CE280 across different fan speeds. But will lose in all but slowest fan speed to 45MM Alphacool XT45. When comparing Alphacool ST30 to XT45, the thinner will lose at all fan speeds. Thermal load is also very important - and here its ok as 180W for 240rads is kinda enough. So practically the answer to the question which rad to pick is kinda more complicated than just thickness. Starting with how much heat should be dissipated at required noise level at the given ambient (some have room temp of 19C while others can easily have around 30C). From there moving to how much space in case and what is required for tube routing and fill ports - thinner larger rad (say 280) will yield better cooling than thicker 240 in any any scenario. Sometimes even things like wideness of the rad can be a factor. For example Alphacool's rads are 1-2mm wider than fans, while HWLabs rads tend to be ~13mm wider. Those could be critical for SFF builds. As for the video topic, I noticed that there is very little cooling improvement in noise optimized systems (fans below 1500) with thicker rads. The biggest advantage is the number of ports. But there are thin rads with multiple ports - Alphacool. Not the best performance, but very compact and have largest number of ports. I'm in a SFF sect for very long time. So I could be wrong about performance in large cases.
Glad somebody wrote this out so I didn’t have to. I usually like this guys videos, but this one needs deleted. The performance metrics aren’t even close to controlled here for the reasons you stated. He should have purchased 3 rads from same company, same series, same FPI, and different thicknesses…. This video will only lead astray the less educated/experienced.
I'm curious to see this test done with more heat being dumped into the system. Your experiment was certainly controlled, but the results are bugging me and I'm definitely interested in how things would change if there was a high power GPU dumping heat into the loop as well.
then a 240mm radiator is not enough since the top party of the case can also support a rad a thin 20-25mm or standard 30mm rad might fit that case well you can either reduce the rad thickness or reduce fan thickness i.e. 15mm thick fans well as long as everything fits.. but if there is a high power GPU then you will need 2x 240mm radiator in that case and i will prefer to use the 45mm thick rad at the bottom part.
well if this is an SFF build and your using the most high end parts like 14900k & 4090 then you want to consider adding a external radiator using a QC fitting and pass through fitting bracket if you don't want to drill holes on the case. for this type of set up a lot recommend a minimum of 3x 360mm radiator a 2x 360mm radiator is fine but at stock settings for both CPU & GPU and you will still running the fans at higher RPM to avoid that you need another 360mm radiator to lower than fan with and even OC .
Thanks to the video, it really saved my money.. I had two heatkiller rad 360S, was thinking about upgrade one of them from S to L (30mm to 48mm). But it seems like it probably would only improve 1C if I do that…
I was wondering earlier this weekend whether running a top and bottom TX240 with 15mm Noctuas in an NR200P would fit. And then I wondered if Machines and More had made any new NR200P videos. I'm not disappointed.
Great content! I've been thinking about a custom loop in my cougar qbx, with dual 240 20,5mm rads. I'm curious to see how slim 15mm fans compare against the 12x25
I would be keen to see the difference on a thicker rad like the EKWB PE with 15 and 25mm fans. E.g. is it better to get more radiator or more fan if your space constrained
More fan and less radiator thickness is always the answer. What most people don't realize is that the majority of performance gained or lost is due to airflow. Thick radiators only make sense if there isn't enough space for more radiator surface, i.e. longer and wider radiators. You CAN cool things with thick radiators, but then you need to have powerful fans running at very high rpm - it's not the ideal situation.
More cooling more better is my motto. To each his own, but I prefer overkill in large or larger case. Everybody and their brother may have one but I love my Lian Li O11DXL-X O11 Dynamic XL ROG, you can squeeze in a 1000W+ of cooling.
Yes, agree with the other commenters that this is a great channel. The concepts are presented clearly and concisely. I was wondering if you've ever worked with Quick Disconnect fittings and their impact on a loop. Thank you
I have not, never saw the value in them for my own use (they’re quite $$ for a set) for it but from my understanding there is typically some added restriction from them, no big deal for a DDC pump or a D5 cranked up a bit more. Now if you’re running something weak like a DC-LT then yeah, there might be a detriment.
@@MachinesMore Thanks for your response. :-) I've got a very dense and small SFF build and every time I need to change a component I have to take out the whole loop in one piece. With quick disconnect fittings I'd be able to remove a single loop component which would be a lot more convenient. My pump is the Swiftech Apogee Drive II which I know is very powerful but I thought I'd ask anyway. Thanks so much!
For a very big rad that say... takes up all your front intakes I'd never go with the thick ones or super-dense high performance fins. Maybe there's a second rad in the loop that can take up that role. Low air impedence is how you make your stuff near silent at low utilization and idle. Where it's gonna be most of the time. The thin rads with high density fins to make up for it... godsend if you need the clearance, but the tradeoff is not great otherwise.
It appears that internal pipe length matters more than anything. Adding more rads would appear to be more beneficial than have a single thicker rad. I wonder if stacking two 20mm thick rads on top of each other would be better then a single 40mm rad. Perhaps do a 240 and a 360mm rad stacked on top of each other instead of a thick 40mm 360mm radiator. That way you can get the hoses to loop properly.
Excelent content, was waiting for a test like this, thanks. Can You run a comparison using noctua thin fans (15mm) thick. That would be a more real comparison, since you normally run the slim rad along with slim fans too un the NR200 for instance
great all I wanted to know was if thickness brought anything because I can go pull fan-54mm but not push fan-54 I'd have to go push fan-30mm because of how my case is, space for 30+30 not 30+54 apparently it does so I'll keep my current pull config for that vertical rad but will go push for the one at the bottom (lianli 011 dynamic xl)
Thank you for this video. i was doubting between a 30mm radiator with slim fans or a 20mm radiator with regular fans. Definitely going for the tx360 now, especially with the extra FPI it should do way better with regular fans.
I've been doing a ton of research on radiator performance and the bottom line is the biggest gains are always in the fan department. Airflow to and out of the radiator is another often overlooked point, as keeping the air cool and moving at a good pace is where the system ultimately succeeds or fails.
Well, that's really surprising. Roughly double the cooling surface area with double the fluid heatsink and there's less than a 2 degree difference? Wow. I would have guessed a much, much greater difference.
Really surprising to see that rad thickness really has very little impact in temps, you can get a bigger reduction in temps by switching to a better thermal compound that you can get from doubling your rad thickness.
in the first graph you didn't let the thermals equalize, you can see the temps are still increasing. in the second graph it looks like it may have just reached it but you should probably have still gone longer.
Question. Which way should I go having place for 360 thick top, 280 mid mid, 360 thin/mid bottom with ~550-600W max power (GPU+CPU) to make it ultra silent in stress? (one of those at least needs to be intake) 1) Populate all and stick to thin/mid ones 2) Dedicate bottom just for unobstructed air intake and get thickets possible mid/top. 3) Doesnt matter - both options can handle. 4) Something else.
Great video, another test i would like to see is how thick a rad needs to be to compensate for length, for instance you don't have space for a 420mm 30mm but can fit a 280 60mm will it be the same performance?
I got a cheap 360x25, the core definitely thinner, closer to 20mm, but I was very surprised to find it cooled my 3080, 320w, without trouble. My main gaming PC has a 4090/13900K, but the rads on that are massive overkill, 3x480’s
The problem with thicker rads is that they usually have lower FPI than thinner ones which in the end leads to not that much of a difference! In the old days you could get Phobya radiators in 6cm thickness and 22 FPI which was insane! Nowadays it's much more boring 16FPI 😢
Nice! I've always wanted to see this test done, and I'm not sure if I had ever seen it done before. It would've been nice to also include a noise comparison, since from what I've heard from the likes of JayzTwoCents is that higher fin density makes for more noise. Maybe a dB comparison wouldn't have been necessary, but at least hear your description of which one sounded less intrusive, or less loud.
I can test since I still have it setup. In my experience, when noise normalized at moderate fans speeds, typical push pull yields a benefit in the neighborhood of 1-2 degrees, putting it on par with a thick rad. The benefit is more that the thermal headroom is much higher if you disregard noise. If you have space for either, I’d personally go thick rad just because the rads are similar in cost, but another set of fans is going to be an added expense, perhaps a splitter cable too, more cables to manage.
@@MachinesMore Do you by chance have extra slim noctua fans (or maybe use ones from top header) to compare thicker rad with slim fans vs thin rad and regular fans? Point is the space in the case is always a limiting factor so with the same space available is it better to go with bigger rad or bigger fan? Or slim rad and slim fan plus leftover space for air flow (once you put that gpu with waterblock on top of that bottom radiator).
Would be curious to see a custom loop NR200P with a thick rad, riser cable and just top mounted fans to see if it's doable. Been eyeing a 54mm r240 but I don't want to make the investment without knowing for sure. Also unclear if you could fit thin fans with something that thick. Can't seem to find GPU slot to bottom of the case clearance anywhere.
You have about 70mm or so because of the power supply cables. If you shift the PSU up or cable manage super super well… the riser cable limits you to 75mm. Either way 54mm is too thick with 25mm fans. I have an XR7 and it absolutely needs 25mm fans, slim fans won’t cut it with that thick rad. EK PE240 is the thickest I would go on the bottom- it was a tight fit already!
great video, thanks for the testing! :) For quite some time i was wondering what the optimal set up would be if you have limted space like in the NR200 below the GPU. Would a normal radiator with slim fans or a thin radiator with normal fans perfome better. Since in your testing the these two rads had similar performance i guess it should be better to grab the thin radiator with regular fans?
I’d go for the TX240 for space reasons. It does actually make good use of its thickness (thinness) by minimizing the offset between the frame and the core. That means you have to be very careful in screws choice but if you use the included screws your should be fine. You’re right in that below the GPU block there’s not much space. With the NR200 dual TX240 loop I opted for slim fans for better airflow, even though 25mm ones would have worked, they would have been right on the block.
45mm starts from a lowe temp... it means that the ambient temp was lower from the beginning. 45m at equal air temp would have performed same as 35mm rad.
Always. PC fans have quite a low performance in reality, compared to what we see on industrial applicationgs, HVAC systems and cars etc. Fans therefore need all the help they can get to cool your water effectively. A radiator will only work well if you can get lots of cool air to flow through it. That's why thick radiators typically have unexpectedly poor performance gains - you just can't effectively cram enough air through them using our little 12V fans. If you want a deep dive, go check out xtremerigs radiator roundup from 2015. It's an old article but they really did a professional job and the physics are spot on. You'll see that not everything we've been taught is factual.
And to clarify further, push only does work with slim radiators, but with thick radiators you need push+pull because the air flow restriction is too much in push only.
You know black tubing is like 4$ a meter on Aliexpress right? So much better looking unless you want Colored liquid ofc... But that gunks up your whole loop though...
Really suprised with the results, I’m planning on water cooling my Jonsbo d30 and been looking at different radiators/fan configs and if it would be better to do a 30mm thick rad with 15mm Arctic p12s or to do a 20mm thick rad with Normal p12
I'm about to build my first custom loop in the NR200P. 240mm slim from EK at the bottom and 92mm on the back. Nothing on top as I'm not comfortable with modding the case yet. Biggest caveat is that my hardware is a 5900x and 3080ti FE. Should I be worried?
It will work, but you will see high fan speeds to keep the coolant temp in check so make sure you are running the best fans you can. With a 92mm rad vertical mounting the card is not an option- with a 30mm rad you will only have space for 15mm slim fans if you want airflow, so Noctua slims. One other thing is the EK Slim rad (guessing you mean the SE 240?) isn’t the best rad in the size class, since you are going to be seeing higher coolant temps, you’re going to want the best possible rad. Of course if you have it already then stick with it, but for the
Late comment but would the variable of pump speed/fluid flowrate be significant. Case in point: I use alphacool DC-LT 2600 tiny pump with rated flowrate of 75L/h in NR200. Know you can faster DDC pumps, so curious.
@ZEFIR001 I typically run Push/Pull and adjust the rpm down for the best noise to cooling ratio I'm guessing a 45mm would really be the sweet spot with a push/pull.
You'll also see a much more significant difference in a system where the custom loop is already being pushed to its limit. Say two high end graphics cards on top of a cpu.
There is a simple thing that can bring temps down. drop the ambient temp. How? coolbot and a common enough window unit ac. 3d print a shroud to an 8 inch vent tube, and pipe the cool air anywhere you want. buy a cheap enough marijuana grow tent, and pipe the cold air into it. Spending 200 on a custom cooling loop just to gain a few extra degrees... AIO's give you at best 44 c temps. custom loops give you at best 37 c temps. With an AC unit cooling the air to 3c, you can expect 20 c temps. Not to mention the positive air pressure, filtered air, and dry air... You could probably get extremely descent temps with just an air cooler like Noctua. Heat pipes really are the most effecient.
Big Chinese brands like bykski and barrow are better than western parts btw. Usually made of copper and half the Price with the same quality. Aluminium is just for cost savings, it's worse In every way... This is one reason why cars use aluminium btw, for those that use that flawed argument against copper. Those radiators are so huge it makes a lot of difference
Your video made this even more important. Buying a better radiator is more important than buying the thickest one. oh and... Having the best fans is a no brainer since you're already spending more on cooling than your actuall Gpu regardless. It needs to be worth it
@@OutOfNameIdeas2imo having more surface area is the best, i went overkill with my setup but I cheated out on fans, I have 2 480mm and 2 360mm, both 60mm thick in push pull (case is a Corsair 1000D) 30 fans (Arctic P12)
Not a great test. You need to test the same line with different thickness. Even if these rads were the same thickness they'd have different performance.
The truth is, the radiators only matter if the temps keep going up. If it ever levels off at a reasonable temperature, you're golden. It's a watts in and watts out equation.
Such an underrated channel. You explain everything in such detail, and with such eloquence. I hope your channel goes to the moon my guy.
Right?! Amazing production quality!
Agreed.
I have a background in heat transfer, just to piggyback that the amount of heat flow gets dissipated into ambient via a rad plateaus really quickly against the fin length (rad thickness in this case), it's actually a tanh curve if I'm not mistaken. In other words, once passing a certain height (which is determined by many factors itself), you are not going to notice any meaningful gains, which is demonstrated in this video. In convection heat exchanger, fin density and length of the rad matters much more. Honestly high pressure head fans at low RPM with low resistance rads works wonder.
This stuff is so fascinating. I’ve been playing around with a custom loop and two 240 rads for the last year and I’m STILL getting surprised at how certain configurations manage heat, air flow etc.
Not to mention the case itself impacts the end result too ❤❤ it’s all one big beautiful complicated, yet understandable, mess 😂
yes, all else being equal, but generally, you have variable core amounts, variable restriction, double stack fins vs single.
I can have seen thinner denser fin stacks lose to thicker less dense fin stacks in bennchmarks at various speeds.
most valuable video about picking a right radiator for your CL .... this was pleasure even to scroll back and listen again and again... dude more of thise dude MORE !
Great channel! Currently researching on switching my NR200, 5800X and 6900XT from air to custom loop.
This is what we need. No BS reviews like the mainstream reviewers. Keep it real.
I've seen this data replicated over the past 2 (or so) years and found it still holds up. In my SFF builds, cooling comes down to the fans. The dirt cheap radiator has always worked better in combination with the Noctua fans. And yes, the pumps are a factor. Indeed poor pump leads to poor results and elevated temps. AIOs are a good example (exit costume loops) since some rpms are "fixed".
Love the info and please keep it going. And, of course, thanks for the content!
Amazing video! Awesome production quality!
Thank you so much for the thorough stuff, concise explanation on findings, accurately representing graphs that look sweet, and the room level speaking volume. My ears and eyes appreciated this video.
This video is fantastic. The visuals helped me understand a lot more then if you were to just say it. Thank you!
Thanks for the great content. Never thought the TX240 would perform this well.
The reason the thin rad underperformed at low fan speeds is because of the higher fin density. That's why the gap shrinks at high fan speeds.
Well presented very interesting topic. I really enjoyed to watch it.
Though I can't totally agree with conclusion.
The thing is that in order to answer the question of rad thickness affect on cooling in a more scientific way, some conditions should be met. And the most important is that the rads should be all from the same series of the same manufacturer. This is because the difference between various models can be huge.
For example the 30mm HWLabbs Nemesis GTS 280 will slightly outperform 45mm EK CoolStream CE280 across different fan speeds.
But will lose in all but slowest fan speed to 45MM Alphacool XT45.
When comparing Alphacool ST30 to XT45, the thinner will lose at all fan speeds.
Thermal load is also very important - and here its ok as 180W for 240rads is kinda enough.
So practically the answer to the question which rad to pick is kinda more complicated than just thickness.
Starting with how much heat should be dissipated at required noise level at the given ambient (some have room temp of 19C while others can easily have around 30C).
From there moving to how much space in case and what is required for tube routing and fill ports - thinner larger rad (say 280) will yield better cooling than thicker 240 in any any scenario.
Sometimes even things like wideness of the rad can be a factor. For example Alphacool's rads are 1-2mm wider than fans, while HWLabs rads tend to be ~13mm wider.
Those could be critical for SFF builds.
As for the video topic, I noticed that there is very little cooling improvement in noise optimized systems (fans below 1500) with thicker rads. The biggest advantage is the number of ports. But there are thin rads with multiple ports - Alphacool. Not the best performance, but very compact and have largest number of ports.
I'm in a SFF sect for very long time. So I could be wrong about performance in large cases.
Thanks to your opinion
Glad somebody wrote this out so I didn’t have to. I usually like this guys videos, but this one needs deleted. The performance metrics aren’t even close to controlled here for the reasons you stated. He should have purchased 3 rads from same company, same series, same FPI, and different thicknesses…. This video will only lead astray the less educated/experienced.
Thanks, this is great encouragement for my tiny PC heart.
I'm curious to see this test done with more heat being dumped into the system. Your experiment was certainly controlled, but the results are bugging me and I'm definitely interested in how things would change if there was a high power GPU dumping heat into the loop as well.
then a 240mm radiator is not enough since the top party of the case can also support a rad a thin 20-25mm or standard 30mm rad might fit that case well you can either reduce the rad thickness or reduce fan thickness i.e. 15mm thick fans well as long as everything fits.. but if there is a high power GPU then you will need 2x 240mm radiator in that case and i will prefer to use the 45mm thick rad at the bottom part.
well if this is an SFF build and your using the most high end parts like 14900k & 4090 then you want to consider adding a external radiator using a QC fitting and pass through fitting bracket if you don't want to drill holes on the case. for this type of set up a lot recommend a minimum of 3x 360mm radiator a 2x 360mm radiator is fine but at stock settings for both CPU & GPU and you will still running the fans at higher RPM to avoid that you need another 360mm radiator to lower than fan with and even OC .
this helped clarify the difference in temps. thanks!!
Thanks to the video, it really saved my money.. I had two heatkiller rad 360S, was thinking about upgrade one of them from S to L (30mm to 48mm). But it seems like it probably would only improve 1C if I do that…
This is really good data.
I was wondering earlier this weekend whether running a top and bottom TX240 with 15mm Noctuas in an NR200P would fit. And then I wondered if Machines and More had made any new NR200P videos. I'm not disappointed.
Great content! I've been thinking about a custom loop in my cougar qbx, with dual 240 20,5mm rads. I'm curious to see how slim 15mm fans compare against the 12x25
I would be keen to see the difference on a thicker rad like the EKWB PE with 15 and 25mm fans. E.g. is it better to get more radiator or more fan if your space constrained
@@inmypaants did you ever find your answer?
More fan and less radiator thickness is always the answer. What most people don't realize is that the majority of performance gained or lost is due to airflow. Thick radiators only make sense if there isn't enough space for more radiator surface, i.e. longer and wider radiators. You CAN cool things with thick radiators, but then you need to have powerful fans running at very high rpm - it's not the ideal situation.
Amazing video! Thanks a lot.
More cooling more better is my motto. To each his own, but I prefer overkill in large or larger case. Everybody and their brother may have one but I love my Lian Li O11DXL-X O11 Dynamic XL ROG, you can squeeze in a 1000W+ of cooling.
Thank you for the content.
Yes, agree with the other commenters that this is a great channel. The concepts are presented clearly and concisely.
I was wondering if you've ever worked with Quick Disconnect fittings and their impact on a loop.
Thank you
I have not, never saw the value in them for my own use (they’re quite $$ for a set) for it but from my understanding there is typically some added restriction from them, no big deal for a DDC pump or a D5 cranked up a bit more. Now if you’re running something weak like a DC-LT then yeah, there might be a detriment.
@@MachinesMore Thanks for your response. :-)
I've got a very dense and small SFF build and every time I need to change a component I have to take out the whole loop in one piece.
With quick disconnect fittings I'd be able to remove a single loop component which would be a lot more convenient.
My pump is the Swiftech Apogee Drive II which I know is very powerful but I thought I'd ask anyway.
Thanks so much!
I’d love to see more rad testing. It’s a topic barely touched on RUclips, except AIOs, which don’t count imo
Very well demonstrated. Thanks! Thickness doesn't seem to be THAT important. Also, choose length over thickness whenever possible.
Looking to cool my 2U (7600X is dumping too much heat for low profile). Sometimes you gotta go smaller (80mm).
I can’t remember if you have done the slim fan/thick rad vs reg fan/slim rad test- based on this the best cooling will likely be on the fan side.
I’ll do some testing for a follow up with slim fans!
For a very big rad that say... takes up all your front intakes I'd never go with the thick ones or super-dense high performance fins.
Maybe there's a second rad in the loop that can take up that role.
Low air impedence is how you make your stuff near silent at low utilization and idle. Where it's gonna be most of the time.
The thin rads with high density fins to make up for it... godsend if you need the clearance, but the tradeoff is not great otherwise.
Thank you! Super helpful
Good job!
It appears that internal pipe length matters more than anything. Adding more rads would appear to be more beneficial than have a single thicker rad. I wonder if stacking two 20mm thick rads on top of each other would be better then a single 40mm rad. Perhaps do a 240 and a 360mm rad stacked on top of each other instead of a thick 40mm 360mm radiator. That way you can get the hoses to loop properly.
Excelent content, was waiting for a test like this, thanks. Can You run a comparison using noctua thin fans (15mm) thick. That would be a more real comparison, since you normally run the slim rad along with slim fans too un the NR200 for instance
Good idea. Are you referring to thin rad with slims?
@@MachinesMore yes, it would be great to know how much the temp change when using the thin rad (20.5 mm) with slim fans
great all I wanted to know was if thickness brought anything because I can go pull fan-54mm but not push fan-54 I'd have to go push fan-30mm because of how my case is, space for 30+30 not 30+54 apparently it does so I'll keep my current pull config for that vertical rad but will go push for the one at the bottom (lianli 011 dynamic xl)
Thank you for this video. i was doubting between a 30mm radiator with slim fans or a 20mm radiator with regular fans. Definitely going for the tx360 now, especially with the extra FPI it should do way better with regular fans.
I've been doing a ton of research on radiator performance and the bottom line is the biggest gains are always in the fan department. Airflow to and out of the radiator is another often overlooked point, as keeping the air cool and moving at a good pace is where the system ultimately succeeds or fails.
Well, that's really surprising. Roughly double the cooling surface area with double the fluid heatsink and there's less than a 2 degree difference? Wow. I would have guessed a much, much greater difference.
Really surprising to see that rad thickness really has very little impact in temps, you can get a bigger reduction in temps by switching to a better thermal compound that you can get from doubling your rad thickness.
in the first graph you didn't let the thermals equalize, you can see the temps are still increasing. in the second graph it looks like it may have just reached it but you should probably have still gone longer.
Question. Which way should I go having place for 360 thick top, 280 mid mid, 360 thin/mid bottom with ~550-600W max power (GPU+CPU) to make it ultra silent in stress? (one of those at least needs to be intake)
1) Populate all and stick to thin/mid ones
2) Dedicate bottom just for unobstructed air intake and get thickets possible mid/top.
3) Doesnt matter - both options can handle.
4) Something else.
Great video.
I'm building a nr200 with a tx240 and I was wondering what I was losing not buying a thicker rad.
I think you will be fine. If you can fit two of those, now you will be very happy.
“What really matters is how you use it”
😆
Great video, another test i would like to see is how thick a rad needs to be to compensate for length, for instance you don't have space for a 420mm 30mm but can fit a 280 60mm will it be the same performance?
I got a cheap 360x25, the core definitely thinner, closer to 20mm, but I was very surprised to find it cooled my 3080, 320w, without trouble.
My main gaming PC has a 4090/13900K, but the rads on that are massive overkill, 3x480’s
The problem with thicker rads is that they usually have lower FPI than thinner ones which in the end leads to not that much of a difference!
In the old days you could get Phobya radiators in 6cm thickness and 22 FPI which was insane! Nowadays it's much more boring 16FPI 😢
Nice! I've always wanted to see this test done, and I'm not sure if I had ever seen it done before. It would've been nice to also include a noise comparison, since from what I've heard from the likes of JayzTwoCents is that higher fin density makes for more noise.
Maybe a dB comparison wouldn't have been necessary, but at least hear your description of which one sounded less intrusive, or less loud.
Which is better: 1 set of fans with a thick rad, or 2 sets of fans doing push-pull on a thin rad?
I can test since I still have it setup. In my experience, when noise normalized at moderate fans speeds, typical push pull yields a benefit in the neighborhood of 1-2 degrees, putting it on par with a thick rad. The benefit is more that the thermal headroom is much higher if you disregard noise. If you have space for either, I’d personally go thick rad just because the rads are similar in cost, but another set of fans is going to be an added expense, perhaps a splitter cable too, more cables to manage.
@@MachinesMore Thank you, that's really helpful!
@@MachinesMore Do you by chance have extra slim noctua fans (or maybe use ones from top header) to compare thicker rad with slim fans vs thin rad and regular fans? Point is the space in the case is always a limiting factor so with the same space available is it better to go with bigger rad or bigger fan? Or slim rad and slim fan plus leftover space for air flow (once you put that gpu with waterblock on top of that bottom radiator).
Corsair's XR 5 and 7 360mm are what I am going with.
Would be curious to see a custom loop NR200P with a thick rad, riser cable and just top mounted fans to see if it's doable. Been eyeing a 54mm r240 but I don't want to make the investment without knowing for sure. Also unclear if you could fit thin fans with something that thick. Can't seem to find GPU slot to bottom of the case clearance anywhere.
You have about 70mm or so because of the power supply cables. If you shift the PSU up or cable manage super super well… the riser cable limits you to 75mm. Either way 54mm is too thick with 25mm fans. I have an XR7 and it absolutely needs 25mm fans, slim fans won’t cut it with that thick rad. EK PE240 is the thickest I would go on the bottom- it was a tight fit already!
great video, thanks for the testing! :)
For quite some time i was wondering what the optimal set up would be if you have limted space like in the NR200 below the GPU.
Would a normal radiator with slim fans or a thin radiator with normal fans perfome better. Since in your testing the these two rads had similar performance i guess it should be better to grab the thin radiator with regular fans?
I’d go for the TX240 for space reasons. It does actually make good use of its thickness (thinness) by minimizing the offset between the frame and the core. That means you have to be very careful in screws choice but if you use the included screws your should be fine. You’re right in that below the GPU block there’s not much space. With the NR200 dual TX240 loop I opted for slim fans for better airflow, even though 25mm ones would have worked, they would have been right on the block.
but my intuition would also tell me that the thicker the rad, the better fan you need to have
so it can push/pull with enough thrust
Very good... but i just worried about the test made with different brands. I mean.. different brands have different projects.
wonder what's the difference between tx240 and gts280 low profile
45mm starts from a lowe temp... it means that the ambient temp was lower from the beginning. 45m at equal air temp would have performed same as 35mm rad.
So, a push+pull with a thin radiator would be better than just push+45mm rad from other tests I've seen around?
Always. PC fans have quite a low performance in reality, compared to what we see on industrial applicationgs, HVAC systems and cars etc. Fans therefore need all the help they can get to cool your water effectively.
A radiator will only work well if you can get lots of cool air to flow through it. That's why thick radiators typically have unexpectedly poor performance gains - you just can't effectively cram enough air through them using our little 12V fans.
If you want a deep dive, go check out xtremerigs radiator roundup from 2015. It's an old article but they really did a professional job and the physics are spot on. You'll see that not everything we've been taught is factual.
And to clarify further, push only does work with slim radiators, but with thick radiators you need push+pull because the air flow restriction is too much in push only.
cool! 2k rpm would be nice to see too
I am curious. Are thick radiators louder?
Noctua fans are pretty ass. Might want to consider replacing those with Phanteks, Thermaltake, Nidec or Scythe fans
Thanks man
You know black tubing is like 4$ a meter on Aliexpress right? So much better looking unless you want Colored liquid ofc... But that gunks up your whole loop though...
Are those fans a little thin for radiators?
You are very likeable, and have some great vids. Subbed
Really suprised with the results, I’m planning on water cooling my Jonsbo d30 and been looking at different radiators/fan configs and if it would be better to do a 30mm thick rad with 15mm Arctic p12s or to do a 20mm thick rad with Normal p12
Always get the thickest possible rad with normal sized fans. Thin fans are gonna be MUCH louder to achieve the same temp
I'm about to build my first custom loop in the NR200P.
240mm slim from EK at the bottom and 92mm on the back.
Nothing on top as I'm not comfortable with modding the case yet.
Biggest caveat is that my hardware is a 5900x and 3080ti FE.
Should I be worried?
It will work, but you will see high fan speeds to keep the coolant temp in check so make sure you are running the best fans you can. With a 92mm rad vertical mounting the card is not an option- with a 30mm rad you will only have space for 15mm slim fans if you want airflow, so Noctua slims. One other thing is the EK Slim rad (guessing you mean the SE 240?) isn’t the best rad in the size class, since you are going to be seeing higher coolant temps, you’re going to want the best possible rad. Of course if you have it already then stick with it, but for the
Late comment but would the variable of pump speed/fluid flowrate be significant. Case in point: I use alphacool DC-LT 2600 tiny pump with rated flowrate of 75L/h in NR200. Know you can faster DDC pumps, so curious.
Everything above 1 GPM is pretty much negligible gains.
guess going 280 slim rad in my case was the way to go
Very Last Post but did you try a push/Pull?
@ZEFIR001 I typically run Push/Pull and adjust the rpm down for the best noise to cooling ratio I'm guessing a 45mm would really be the sweet spot with a push/pull.
Wow, so only about 10-15% difference between the 20mm TX240 and the 45mm EK
Any chance you can tell us what was your CPU package power when you were conducting these tests?
Just under 180W.
thank you
You'll also see a much more significant difference in a system where the custom loop is already being pushed to its limit. Say two high end graphics cards on top of a cpu.
dang that rad is THICC
There is a simple thing that can bring temps down.
drop the ambient temp.
How?
coolbot and a common enough window unit ac. 3d print a shroud to an 8 inch vent tube, and pipe the cool air anywhere you want.
buy a cheap enough marijuana grow tent, and pipe the cold air into it.
Spending 200 on a custom cooling loop just to gain a few extra degrees...
AIO's give you at best 44 c temps.
custom loops give you at best 37 c temps.
With an AC unit cooling the air to 3c, you can expect 20 c temps. Not to mention the positive air pressure, filtered air, and dry air...
You could probably get extremely descent temps with just an air cooler like Noctua. Heat pipes really are the most effecient.
My wife says thickness doesn't matter.
I have doubts though.
Any p[lan on reviewing the SSUPD Meshilicious?
😉
Watching this with my 360mmx38mm Arctic Liquid Freezer 3
So, it really doesn't matter all that much. Especially not with multiple rads.
Two thin rads when placed at intake points will way outperform one double thick rad.
Big Chinese brands like bykski and barrow are better than western parts btw. Usually made of copper and half the Price with the same quality. Aluminium is just for cost savings, it's worse In every way... This is one reason why cars use aluminium btw, for those that use that flawed argument against copper. Those radiators are so huge it makes a lot of difference
Your video made this even more important. Buying a better radiator is more important than buying the thickest one. oh and... Having the best fans is a no brainer since you're already spending more on cooling than your actuall Gpu regardless. It needs to be worth it
@@OutOfNameIdeas2imo having more surface area is the best, i went overkill with my setup but I cheated out on fans, I have 2 480mm and 2 360mm, both 60mm thick in push pull (case is a Corsair 1000D) 30 fans (Arctic P12)
i thought thick rads were for soak time not instant cooling affects..........
T H I C C
Not a great test. You need to test the same line with different thickness. Even if these rads were the same thickness they'd have different performance.
The truth is, the radiators only matter if the temps keep going up. If it ever levels off at a reasonable temperature, you're golden. It's a watts in and watts out equation.
Thicc with two c