Great video! Very informative and full of useful data. Covers the most popular setups for sure. Cpu alone produces probably around 100watts. Gpu probably doubles that. My take away is that for 1100 fan speed triple rad is the minimum. The testing must have taken ages. Thanks for the effort
Great video! I have never seen a video with this experiment/comparation and I think as you say it is a very common question for water cooling starters, thanks for the info, keep it up!
Awesome vid man, after coming up to near completion of my first ridiculous loop, I think my next loop will be to use all of the stuff I learned to make a loop using no more than is necessary, following the KISS principle, that will work in any atx case without any cable management needed. Definitely good advice, don't use more radiator than is needed!
Good comparison, though I would say that certain radiators with different flow rates seem to perform better @ higher or lower RPM like the Alphacool Monsta radiators seem to perform better @ lower rpm fan speeds.
Very interesting, now I don't have to worry about my plan for a 140x2 Rad, for a 5930K and a R9 290X (with no overlocking) then an additional 140x2 when I get a second gfx card. I'm going to be working in a Mercury S8, and I'm aiming to ran as few fans as possible, as slowly as possible.
always get the biggest one you can fit. they cost nearly the same either way and you get alot more adding another "fan space" rather than thickness, fpi and fan speed. it is THE most silent way to get cooler temps.
Awesome video. Keep up the good work. You answered a question I often wondered about. I am about to build my first custom water cooled system and I was unsure if I should cram every space I could with radiators. This video has given me some things to think about.
I really prefer 240mm per part and high mm-h2o as my cooler at 2500rpm only has mm--h20 of 2.38 when many after market fans have that at lower speeds So i say invest in great after market fans
Nice. Just getting into this water cooling thing and it is confirmed that I was on point just going with a 360 for my cpu and 980 Ti. Now if I can just decided which pwm fan to get. Thanks for the video - liked
Thank you for the results. I see people here suggesting that this is not very representative, as a 290x would be bad in your configuration since it runs hot, but really, is it bad? As far as I know, using a AIO single 120 rad with a mounting bracket such as G10 on a 290x usually results in amazing temperatures. And furthermore, a 295x2 runs two 290x gpu's on a single 120 rad. So no, I really do not think this test is off. You can easily run an x99 CPU with a 290x/980 from a single 280 rad. That is the same as running two 120 AIO coolers on each component. It is possible. Adding more rad will only benefit so much, as DazMode results show. Once again, this video was very useful!
1500 RPM is OK for you? With what type of Fans, cause I run NB e-loop B12-PS's with max 1050-1100 until I can hear the air moving but most of the time they run with 400-500 RPM which is complete silence. The only thing that is annoying me is sometime when I render to an HDD the sidepanel of the air 540 starts to vibrate, making noise..
Hi, can you make a test like this- thick, low fpi radiator in push or pull vs thin, high fpi in push-pull ? This would be an interesting comparison, because i think you might get better temps and lower noise with cheaper, thinner radiator in push-pull.
Wish you would have tested hot products like fx series cpu and amd gpus to get a real heat test. my system gets hot with my fx8320 overclocked to 5.2ghz.
may i suggest that you start testing noise levels in two different cases - one sound optimized like an r5 and one normal like an enthoo pro. also, since most people will undervolt fans, you should undervolt yours and include the numbers as well. this way, you will get the best and most representative results
Ya so this is complete Bull, I am an "AMD user" as I seen you say elsewhere in the comments. However I have my rig torn down atm to change to a different case. Anyway I went from 2 480s down to 1, the result is my gpus in 70s and my CPU also in the 70s. under a gaming load, thats a change from CPU never passing 50 and my gpus never passing 40. That is a 20 degree difference on CPU and a 30 degree difference on GPUs. My specs if your curious is a I7 4930k Overclock to 4.6ghz 1.45v, and 2 stock clocked atm 290xs however with the other 480 and lower temps they were overclocked to 1175/1600 with 1.4v. The air coming from the 480 on my bench is physically hot very hot that isn't the case with the 2 480s hooked up. This is most likely true if your using not flagship cards and not flagship CPUs that are not overclocked. However higher end stuff with mass overclocks this is no longer true.
i do run my oc'ed 8320 with a 420 rad to niceness to this is that i dont have to run my fans at all when doing mondaine tasks and light gaming, and my cpu gets to 48c in HWMonitor. (i have a graphite t600 so the rad is custom on top of case)
running i5 4670k 2x 290s and wow, they dump heat. with 2x 360 rads I see water temp delta going close to 25C if the whole system is being stressed. Sim city seems to do it well
Very interesting video, 1 question, were the cpu and gpu stock or overcloked? Based on the first diagram as the temps were almost the same, they are stock...?
Hey I'm running ryzen 7 1700x and asus posiedon 1080ti and I'm planning to overclock both mildly. How much radiator would I need? I'm thinking either 1 420 or 3 280?
I want you to use a 240 rad for a gpu and cpu loop and overclock both to the max and see what happens lol that 120mm per item is only at stock clocks buddy.
With that GPU it looks like you need 360 RAD if you play games. Now what's happens when you put a Core i7, and 780 Ti or 290X. Then you will need at least 480. The 120 per part really only apples for mid Range parts in TDP. GTX 770 is not a good example. Most people that go water have much higher end GPUs and more then 1. Also it would be nice to measure water delta temps. I have 3770K @ 4.6GHz and 2 x 290X OCed and can hit 15-20C delta with 360 and 240 thick push pull 1000 rpm.
water temps are irrelevant since it always reaches an equilibrium within minutes of any run. and second, the 770 is a great baseline for this sort of testing. congrats on having dual 290X ... but it is impossible for Daz to measure all variants.
I agree with orjon here, 120mm per hot component is good for mid tdp components and lower power tasks, but gaming on a 290 (especially oc'd) is going to exceed that rule. I have a 4790k @ 4.7 and 2x 290 oc'd under a triple and dual darkside radiator, with SP 120s on max (thank God for IEM's while gaming) I hit 61 on the gpus and 70 on the cpu in dragon age inquisition . If daz gets GTs I may give those a whirl, but I'm pretty sure a quad would struggle to keep up. The reason I got into WC was more for the performance potential and I understand there is the crowd that wants quiet systems at lower power, but us gamers put out more heat :) or at least those of us on 290s. On the bright side, it's been -40 in Toronto lately, but my room is never cold while gaming!
Vu Nguyen I was wondering about performance too. I have a 360 and was concerned it wouldn't keep my gpu and cpu cool enough. I am very surprised it worked much better than I expected. I am running a i7 930 @ 4 ghz/1,27 v and a 290x Lightning @ 1200 core and 1500 memory/1,39 v. I am using 1 loop with EKWB XTX 360 (65 mm copper) and 3 Noctua NF-F12 spinning at 1000 rpm. Torture (Prime95 and Fur) will give me 59c on GPU and 71c on CPU. With ambient at 22c. Offcourse this is a rather high delta, but at normal gaming use I never go above 52c on GPU and 55c on CPU. I will add some rads in the future for lower delta, but this setup runs cool enough for 24/7 use and super quiet.
Well your GPU is probably doing ~ 400-450W alone. Thats like running 3 x GTX970 lol. The CPU is also the least efficient Core i7 also using ~ 200W under load.
Dude, room temparatur maybe at 22 degree and your system is at 33, meaning the delta is 11 degree. You expect this to get lower by 5 degree only by adding some radiators? Thats pure physics man, you cant wish for 50% lower cooling by just adding a little bit of radiator mass. But nice video!
I have a 2500k and a GTX 480. When I am idling (just browsing the internet, or very light gaming) then I can go passive on a dual rad. normal gaming requires one set of fans to be on. Heavy gaming or F@H requires push pull. However I have my fans on low. 1600RPM.
Next to impossible. You will always need so airflow for the air to move through the RAD. For example if you run just the pump though a loop with no fans and system off the water will get hot. D5 pump is cooled by the water but the pump alone is enough to get the water warm which is like 15w. Dump 200w+ and it will burn. I once had a problem with fan controller so all fans in Rad where off. Water in the RAD got so hot it would burn you. System just shut off after that. Only way you can have something possible is if you have case fans and only exist is thought the RAD. The story here is 500 Rpm fans better then 0 rpm. Just look at laptops. Only 10w parts get passive cooling. If noise is the problem then water cooling is not the answer. The pump makes note noise then my 10 AP-15 @ 7v.
I am thinking of having 2 x560s a 120 a 480 and a 420 without any fans. I am wondering if I can get away from needing a 560 and the 420 or not. I the system will have 2-3 next gen amd 300 series gpus (mid to high end) and intel Skylake processor. Maybe the north bridge water cooled. Now I got 1 fx 8350 and a gtx titan.
For those wondering he's putting around 350-400 TDP into his loop.
Man I really love your videos, things that you cant see anywhere else, thanks!
Looks like a 360rad is a nice spot for cpu and single gpu. Thanks for this, cleared up a lot of questions I had.
Great video! Very informative and full of useful data. Covers the most popular setups for sure.
Cpu alone produces probably around 100watts. Gpu probably doubles that.
My take away is that for 1100 fan speed triple rad is the minimum.
The testing must have taken ages. Thanks for the effort
Finally some solid evidence! Great vid Daz! Thanks!
Great video! I have never seen a video with this experiment/comparation and I think as you say it is a very common question for water cooling starters, thanks for the info, keep it up!
Thanks a lot for a great, informative video, Daz!
Awesome vid man, after coming up to near completion of my first ridiculous loop, I think my next loop will be to use all of the stuff I learned to make a loop using no more than is necessary, following the KISS principle, that will work in any atx case without any cable management needed. Definitely good advice, don't use more radiator than is needed!
Good comparison, though I would say that certain radiators with different flow rates seem to perform better @ higher or lower RPM like the Alphacool Monsta radiators seem to perform better @ lower rpm fan speeds.
*Exactly how much performance does buying more radiators get you?*
Very interesting, now I don't have to worry about my plan for a 140x2 Rad, for a 5930K and a R9 290X (with no overlocking) then an additional 140x2 when I get a second gfx card.
I'm going to be working in a Mercury S8, and I'm aiming to ran as few fans as possible, as slowly as possible.
Encountered your channel not long ago, very interesting content, keep it going.
always get the biggest one you can fit. they cost nearly the same either way and you get alot more adding another "fan space" rather than thickness, fpi and fan speed. it is THE most silent way to get cooler temps.
Good info for sure Daz, Thanks
Great vid Daz
Really interesting. Thanks for posting.
Awesome video. Keep up the good work. You answered a question I often wondered about. I am about to build my first custom water cooled system and I was unsure if I should cram every space I could with radiators. This video has given me some things to think about.
I really prefer 240mm per part and high mm-h2o as my cooler at 2500rpm only has mm--h20 of 2.38 when many after market fans have that at lower speeds So i say invest in great after market fans
Nice. Just getting into this water cooling thing and it is confirmed that I was on point just going with a 360 for my cpu and 980 Ti. Now if I can just decided which pwm fan to get.
Thanks for the video - liked
Thank you for the results. I see people here suggesting that this is not very representative, as a 290x would be bad in your configuration since it runs hot, but really, is it bad? As far as I know, using a AIO single 120 rad with a mounting bracket such as G10 on a 290x usually results in amazing temperatures. And furthermore, a 295x2 runs two 290x gpu's on a single 120 rad. So no, I really do not think this test is off. You can easily run an x99 CPU with a 290x/980 from a single 280 rad. That is the same as running two 120 AIO coolers on each component. It is possible. Adding more rad will only benefit so much, as DazMode results show. Once again, this video was very useful!
1500 RPM is OK for you? With what type of Fans, cause I run NB e-loop B12-PS's with max 1050-1100 until I can hear the air moving but most of the time they run with 400-500 RPM which is complete silence. The only thing that is annoying me is sometime when I render to an HDD the sidepanel of the air 540 starts to vibrate, making noise..
stuk some paper betveen the door and try :D
put it this way buy 10x 560mm 86mm thick rads and a very powerful pump to pump the coolant in and out
Hi, can you make a test like this- thick, low fpi radiator in push or pull vs thin, high fpi in push-pull ?
This would be an interesting comparison, because i think you might get better temps and lower noise with cheaper, thinner radiator in push-pull.
Do you have a video comparing tour rads vs other brands?
The rule of thumb is, generally, 120mm per water-cooled component + 120mm to spare.
Hi Daz
is it possible/practical to have enough radiation surface area to cool the system passively (no fans) for a quite system?
Wish you would have tested hot products like fx series cpu and amd gpus to get a real heat test. my system gets hot with my fx8320 overclocked to 5.2ghz.
may i suggest that you start testing noise levels in two different cases - one sound optimized like an r5 and one normal like an enthoo pro. also, since most people will undervolt fans, you should undervolt yours and include the numbers as well. this way, you will get the best and most representative results
Does it matter if a radiator had more outlets?
Ya so this is complete Bull, I am an "AMD user" as I seen you say elsewhere in the comments. However I have my rig torn down atm to change to a different case. Anyway I went from 2 480s down to 1, the result is my gpus in 70s and my CPU also in the 70s. under a gaming load, thats a change from CPU never passing 50 and my gpus never passing 40. That is a 20 degree difference on CPU and a 30 degree difference on GPUs.
My specs if your curious is a I7 4930k Overclock to 4.6ghz 1.45v, and 2 stock clocked atm 290xs however with the other 480 and lower temps they were overclocked to 1175/1600 with 1.4v. The air coming from the 480 on my bench is physically hot very hot that isn't the case with the 2 480s hooked up.
This is most likely true if your using not flagship cards and not flagship CPUs that are not overclocked. However higher end stuff with mass overclocks this is no longer true.
i do run my oc'ed 8320 with a 420 rad to niceness to this is that i dont have to run my fans at all when doing mondaine tasks and light gaming, and my cpu gets to 48c in HWMonitor. (i have a graphite t600 so the rad is custom on top of case)
thank you hehe puter not done yet lol
Great video.
running i5 4670k 2x 290s and wow, they dump heat. with 2x 360 rads I see water temp delta going close to 25C if the whole system is being stressed. Sim city seems to do it well
Thank you so much for this video, interesting results though :$
Very interesting video, 1 question, were the cpu and gpu stock or overcloked? Based on the first diagram as the temps were almost the same, they are stock...?
Both overclocked.
Would 3x 80mm*160mm radiators enough for cooling a cpu and gpu?
Nice T-shirt :D.
Hey I'm running ryzen 7 1700x and asus posiedon 1080ti and I'm planning to overclock both mildly. How much radiator would I need? I'm thinking either 1 420 or 3 280?
What are those temps? Cpu? delta ambient? Liquid?
+Keven Harvey Average cores vs room temp delta.
DazMode
OK, thanks.
2000 Degrees! :D LOL
That's a great shirt.
I want you to use a 240 rad for a gpu and cpu loop and overclock both to the max and see what happens lol that 120mm per item is only at stock clocks buddy.
+Theorcman2008 Usually such comments come from AMD users.
+DazMode yeah? Well I run intel and nvidia.
+Theorcman2008 You should be ok with decent fans and 60mm rad no matter what then.
With that GPU it looks like you need 360 RAD if you play games. Now what's happens when you put a Core i7, and 780 Ti or 290X. Then you will need at least 480. The 120 per part really only apples for mid Range parts in TDP. GTX 770 is not a good example. Most people that go water have much higher end GPUs and more then 1. Also it would be nice to measure water delta temps. I have 3770K @ 4.6GHz and 2 x 290X OCed and can hit 15-20C delta with 360 and 240 thick push pull 1000 rpm.
water temps are irrelevant since it always reaches an equilibrium within minutes of any run. and second, the 770 is a great baseline for this sort of testing. congrats on having dual 290X ... but it is impossible for Daz to measure all variants.
the 290X is a VERY how card, so a pair of them are likely to benefit from additional radiators.
I agree with orjon here, 120mm per hot component is good for mid tdp components and lower power tasks, but gaming on a 290 (especially oc'd) is going to exceed that rule. I have a 4790k @ 4.7 and 2x 290 oc'd under a triple and dual darkside radiator, with SP 120s on max (thank God for IEM's while gaming) I hit 61 on the gpus and 70 on the cpu in dragon age inquisition . If daz gets GTs I may give those a whirl, but I'm pretty sure a quad would struggle to keep up.
The reason I got into WC was more for the performance potential and I understand there is the crowd that wants quiet systems at lower power, but us gamers put out more heat :) or at least those of us on 290s.
On the bright side, it's been -40 in Toronto lately, but my room is never cold while gaming!
Vu Nguyen I was wondering about performance too. I have a 360 and was concerned it wouldn't keep my gpu and cpu cool enough. I am very surprised it worked much better than I expected.
I am running a i7 930 @ 4 ghz/1,27 v and a 290x Lightning @ 1200 core and 1500 memory/1,39 v.
I am using 1 loop with EKWB XTX 360 (65 mm copper) and 3 Noctua NF-F12 spinning at 1000 rpm. Torture (Prime95 and Fur) will give me 59c on GPU and 71c on CPU. With ambient at 22c. Offcourse this is a rather high delta, but at normal gaming use I never go above 52c on GPU and 55c on CPU.
I will add some rads in the future for lower delta, but this setup runs cool enough for 24/7 use and super quiet.
Well your GPU is probably doing ~ 400-450W alone. Thats like running 3 x GTX970 lol. The CPU is also the least efficient Core i7 also using ~ 200W under load.
Dude, room temparatur maybe at 22 degree and your system is at 33, meaning the delta is 11 degree. You expect this to get lower by 5 degree only by adding some radiators? Thats pure physics man, you cant wish for 50% lower cooling by just adding a little bit of radiator mass. But nice video!
us crazy gamers ........ lol
That's with an Intel cpu... Try this test with AMD
The only difference there would be higher loads. Same principle applies
www.dazmode.com/store/product/ek-coolstream-we-360-dual-180mm-radiator/
tes test test :P
Автор говорит по-русски, ведь так?) просто вопрос так удобнее задать.
At what amount of radiator would you need to be able to run the system completely passive. No fans at all.
I have a 2500k and a GTX 480. When I am idling (just browsing the internet, or very light gaming) then I can go passive on a dual rad.
normal gaming requires one set of fans to be on. Heavy gaming or F@H requires push pull. However I have my fans on low. 1600RPM.
Next to impossible. You will always need so airflow for the air to move through the RAD. For example if you run just the pump though a loop with no fans and system off the water will get hot. D5 pump is cooled by the water but the pump alone is enough to get the water warm which is like 15w. Dump 200w+ and it will burn. I once had a problem with fan controller so all fans in Rad where off. Water in the RAD got so hot it would burn you. System just shut off after that. Only way you can have something possible is if you have case fans and only exist is thought the RAD. The story here is 500 Rpm fans better then 0 rpm. Just look at laptops. Only 10w parts get passive cooling. If noise is the problem then water cooling is not the answer. The pump makes note noise then my 10 AP-15 @ 7v.
Try it. First thing to go will be the pump since it has no protection above 65C water temp.
I am thinking of having 2 x560s a 120 a 480 and a 420 without any fans. I am wondering if I can get away from needing a 560 and the 420 or not. I the system will have 2-3 next gen amd 300 series gpus (mid to high end) and intel Skylake processor. Maybe the north bridge water cooled. Now I got 1 fx 8350 and a gtx titan.
***** Same here. With quad temps keep creeping up slowly but surely. I stopped at water 50c. I have to have some air movement through fins.