-
Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!
And voila!!
Nice! I'm curious what's the culprit for both sides bulge a bit? I'm looking to implement this once I received my T1!
Here a few CPU and GPU tests!
The side fans are INTAKING and the top fans are EXHAUSTING.
I will try tomorrow a full NEGATIVE pressure build.
It's interesting to see the difference between the two CPU tests, with the side fans ON and OFF!!
CPU TEST after 30 minutes of Blender rendering, all fans at 100%
CPU - 63C
GPU - 32C
SSDs - 43C (OS Front), 41C (Back)
VRM - 60C
CHIPSET - 57.3C
AMBIENT - 21C
WATER TEMP GPU OUT - 30.7C
DELTA WATER TEMP - 9.7C
-
GPU Test under Heaven 4.0 Benchmark for 30 minutes, all fans 100%
CPU - 67C AVERAGE
GPU - 50C
SSDs - 49C (OS Front), 47C (Back)
VRM - 52C
CHIPSET - 61C
AMBIENT - 20,3C
WATER TEMP GPU OUT - 39.7C
DELTA WATER TEMP - 19.4C
-
CPU TEST after 30 minutes of Blender rendering, all fans at 100%, SIDE FANS OFF
CPU - 66C
GPU - 35C
SSDs - 41C (OS Front), 39C (Back)
VRM - 58C
CHIPSET - 53.8C
AMBIENT - 20,6C
WATER TEMP GPU OUT - 33.6C
DELTA WATER TEMP - 13C
The side fans are INTAKING and the top fans are EXHAUSTING.
I will try tomorrow a full NEGATIVE pressure build.
It's interesting to see the difference between the two CPU tests, with the side fans ON and OFF!!
CPU TEST after 30 minutes of Blender rendering, all fans at 100%
CPU - 63C
GPU - 32C
SSDs - 43C (OS Front), 41C (Back)
VRM - 60C
CHIPSET - 57.3C
AMBIENT - 21C
WATER TEMP GPU OUT - 30.7C
DELTA WATER TEMP - 9.7C
-
GPU Test under Heaven 4.0 Benchmark for 30 minutes, all fans 100%
CPU - 67C AVERAGE
GPU - 50C
SSDs - 49C (OS Front), 47C (Back)
VRM - 52C
CHIPSET - 61C
AMBIENT - 20,3C
WATER TEMP GPU OUT - 39.7C
DELTA WATER TEMP - 19.4C
-
CPU TEST after 30 minutes of Blender rendering, all fans at 100%, SIDE FANS OFF
CPU - 66C
GPU - 35C
SSDs - 41C (OS Front), 39C (Back)
VRM - 58C
CHIPSET - 53.8C
AMBIENT - 20,6C
WATER TEMP GPU OUT - 33.6C
DELTA WATER TEMP - 13C
Last edited:
so 3c with side fans off, do you remember a similar test what cpu was with just the single Rad?
I've been searching reddit and the forum for some answers around the 240 AIO that are 27mm in height like the Corsair h100i. What are peoples experience when mounting the fans on top of the radiator, does it cause any bulging? Is it as cool as having them underneath? I know mounting them underneath has seen some rubbing against the motherboard. Given the clearance is 42mm and the radiator (27mm) and slim fan (15mm) comes in at 42mm I'm worried this could cause problems such bowing if everything isn't perfect. Keen to hear what peoples experiences are, is this something I should be worried about? I've heard some issues with the LT240 which required cleaning, is that still a problem or should I go with that radiator because it's slightly shorter? Thanks!
Now that is an interesting thought, a SFF case where the Rad is part of the case. . the price tag would be super high though, and considering the price of these 'high end' SFF cases already yikes i can only imagine, but sounds interesting for sure though.
Putting the fans above the rad won't make the top panel bulge. The only reason 27mm rads and 15mm fans bulge is because some people (like me) have to use washers to space the fan from the mobo, which adds thickness to the entire unit.
Thermals should be a few degrees warmer (I think I remember some testing from a while ago on this in this thread), and in my very quick testing, the noise levels were a bit louder in terms of air noise.
Thanks Navic (love you build by the way), that's good to know, maybe I'll stick with Corsair h100i and see how I go. I'll have a look around as well to see if I can find any running fans on top, given I'm planning to run a 5600 or 5700 Ryzen 3 I don't think thermals will be too bad with fans on top, hopefully the noise doesn't bother me too much... If it makes a difference I've bought the Gigabyte B550i, do you know if that would have the same issues as your board? I would assume so given the boards are the same size?
There's no way to say definitively. People haven't had the problem with the same board and cooler I have, others have had the same problem with different motherboards and coolers. Because of the tolerances of everything, it's impossible to say if you will or wont have the same problem I had, but I would say that there is a higher (but not 100%) chance that you will not have the same problem I had.
So, a few more tests and news!
I've swapped the 12mm fans for the Scythe KazeFlex Slim 15mm.
Here you can see the clearance left AND WHY the XSPC was not working for me in this configuration, BUT the EKWB SE Classic will work as well instead.
Following the tests with the new fans!
CPU TEST after 30 minutes of Blender rendering, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 63C
GPU - 32C
SSDs - 44C (OS Front), 42C (Back)
VRM - 60C
CHIPSET - 57.3C
AMBIENT - 21,4C
WATER TEMP GPU OUT - 30.4C
DELTA WATER TEMP - 9C
-
CPU TEST after 30 minutes of Blender rendering, OC ALL CORES 4.3GHz 1.25V VCORE, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 68C
GPU - 32C
SSDs - 45C (OS Front), 43C (Back)
VRM - 64C
CHIPSET - 58C
AMBIENT - 21,3C
WATER TEMP GPU OUT - 31.7C
DELTA WATER TEMP - 10.4C
-
@DrLeroy my latest previous test with the GTS 240, TOP Noctuas 15mm, FRONT 120 INTAKING SLIM NOCTUA on the CPU side, AND OC 4.3GHz ON ALL CORES HAD THESE RESULTS
CPU TEST after 30 minutes of Blender rendering, OC ALL CORES 4.3GHz 1.25V VCORE, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 72C
GPU - 40C
SSDs - 45C (INTEL OPTANE), 41C (Front)
VRM - 58C
CHIPSET - 57C
AMBIENT - 23,3C
WATER TEMP GPU OUT - 39C
DELTA WATER TEMP - 15.7C
I've swapped the 12mm fans for the Scythe KazeFlex Slim 15mm.
Here you can see the clearance left AND WHY the XSPC was not working for me in this configuration, BUT the EKWB SE Classic will work as well instead.
Following the tests with the new fans!
CPU TEST after 30 minutes of Blender rendering, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 63C
GPU - 32C
SSDs - 44C (OS Front), 42C (Back)
VRM - 60C
CHIPSET - 57.3C
AMBIENT - 21,4C
WATER TEMP GPU OUT - 30.4C
DELTA WATER TEMP - 9C
-
CPU TEST after 30 minutes of Blender rendering, OC ALL CORES 4.3GHz 1.25V VCORE, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 68C
GPU - 32C
SSDs - 45C (OS Front), 43C (Back)
VRM - 64C
CHIPSET - 58C
AMBIENT - 21,3C
WATER TEMP GPU OUT - 31.7C
DELTA WATER TEMP - 10.4C
-
@DrLeroy my latest previous test with the GTS 240, TOP Noctuas 15mm, FRONT 120 INTAKING SLIM NOCTUA on the CPU side, AND OC 4.3GHz ON ALL CORES HAD THESE RESULTS
CPU TEST after 30 minutes of Blender rendering, OC ALL CORES 4.3GHz 1.25V VCORE, all fans at 100%, SCYTHE KAZEFLEX SLIM 15mm ON THE SIDE
CPU - 72C
GPU - 40C
SSDs - 45C (INTEL OPTANE), 41C (Front)
VRM - 58C
CHIPSET - 57C
AMBIENT - 23,3C
WATER TEMP GPU OUT - 39C
DELTA WATER TEMP - 15.7C
Last edited:
Finished my own dual 240 build!
Everything is a tight fit and I mean everything.
CPU side: Thanks to aquanaut, I was able to install a DDC pump with heatsink in the CPU slot. It currently touches the side panel but there's no flexing. I used quick disconnects feeding directly into the pump and a small ext res to fill the loop.
GPU side: I've installed the 17mm thick rad from taobao along with the 12mm Scythe PWM slim fans to maximise space and there's approximately 10mm of space between the fan and the 2080ti waterblock. Squeezed in the space behind the rad, beside the GPU is an Aquacomputer quadro, powering all four fans, a temp sensor and the pump that I rewired into a single fan connector. All power cabling is a set from pslate customs as there's simply no spare space for excess cabling.
Testing: I ran smallFFT tests on Prime95 as well as Heaven benchmark at same time. CPU is 3700X and GPU is 2080ti at 900mV. Previously installed pump was from the LT240. Total power draw from the CPU and GPU is around 340W. Graphs of the testing can be found here, but a small table of results below. room temp was 22C.
The coolant temp doesn't change too much but there's a large drop in GPU and CPU temps. The CPU is thermal throttling in the 1x240mm test as it's hit the 95 degree mark. Looking at fabio's results, it honestly impresses me how much better the silicon is on the higher end chips. I did further tests with different pump speeds. It's a little audible at 3000rpm BUT is what I'll ramp up to from 2000rpm when under load. At full pump speed I get maximum of 45 degrees on a 2080ti! Massive 21 degree drop from 66 in the single rad setup and 12 degrees off the CPU while 10W more power due to the thermal throttling is also nice.
Everything is a tight fit and I mean everything.
CPU side: Thanks to aquanaut, I was able to install a DDC pump with heatsink in the CPU slot. It currently touches the side panel but there's no flexing. I used quick disconnects feeding directly into the pump and a small ext res to fill the loop.
GPU side: I've installed the 17mm thick rad from taobao along with the 12mm Scythe PWM slim fans to maximise space and there's approximately 10mm of space between the fan and the 2080ti waterblock. Squeezed in the space behind the rad, beside the GPU is an Aquacomputer quadro, powering all four fans, a temp sensor and the pump that I rewired into a single fan connector. All power cabling is a set from pslate customs as there's simply no spare space for excess cabling.
Testing: I ran smallFFT tests on Prime95 as well as Heaven benchmark at same time. CPU is 3700X and GPU is 2080ti at 900mV. Previously installed pump was from the LT240. Total power draw from the CPU and GPU is around 340W. Graphs of the testing can be found here, but a small table of results below. room temp was 22C.
| Test | CPU temp | GPU temp | Coolant temp | Notes |
| 1 x 240mm | 95 | 66 | 42 | LT240 pump + rad |
| 2 x 240mm, 2000rpm | 88 | 51 | 42 | EK-DDC 3.2 PWM pump |
| 2 x 240mm, 3000rpm | 86 | 49 | 41 | EK-DDC 3.2 PWM pump |
| 2 x 240mm, 4000rpm | 84 | 47 | 38.5 | EK-DDC 3.2 PWM pump |
| 2 x 240mm, 4500rpm | 83 | 45 | 37 | EK-DDC 3.2 PWM pump |
The coolant temp doesn't change too much but there's a large drop in GPU and CPU temps. The CPU is thermal throttling in the 1x240mm test as it's hit the 95 degree mark. Looking at fabio's results, it honestly impresses me how much better the silicon is on the higher end chips. I did further tests with different pump speeds. It's a little audible at 3000rpm BUT is what I'll ramp up to from 2000rpm when under load. At full pump speed I get maximum of 45 degrees on a 2080ti! Massive 21 degree drop from 66 in the single rad setup and 12 degrees off the CPU while 10W more power due to the thermal throttling is also nice.
@Van0 awesome! build! I'm gonna go the dual rad route, soon-ish haha!
btw did you measure how tall was the entire pump block combo? I have the Bitspower version of DDC pump with integrated heatsink and it came to 55mm from cold plate to top of the pump so I'm unsure if it would bulge the side panel or not in 3 slot mode.
btw did you measure how tall was the entire pump block combo? I have the Bitspower version of DDC pump with integrated heatsink and it came to 55mm from cold plate to top of the pump so I'm unsure if it would bulge the side panel or not in 3 slot mode.
The EK Coolstream SE 280 (as an example) is 10mm shorter than the T1, and only 7mm wider. The ends are fairly squared off, too, so it'd make a great top. Could use the same rad on the side, which would make it a bit wider than the T1, or could use one of the thinner 240 mm rads like @fabio and @Van0 to maintain a similar case width. Dunno, maybe there's no point - just use two thinner 240mm rads, stick them in the T1, and be done with it.
Thanks for sharing all great builds and insight, it’s been so helpful. I had a question on LT240 vs Corsair h100i, for someone who has never built a water cooled PC before what would be the recommendation? Cost isn’t the concern, it’s more around reliability, maintenance and ease of set up. It seems the LT240 is a favourite and ideally that’s what I would like to go with but do want to know how it compares to Corsair in the areas I mentioned and is it recommended for a newbie. Thank you!
I did, it came to a hair under 54mm. hoping your build has a little extra tolerance so you can fit without a bulge but if my build has taught me anything, 1mm can be the difference. Eg I had to remove a screw from my A12x25 just to squeeze the quick disconnects under.
yes, 1mm is HUGE in the SFF world I know first hand from all the cable bending and stuffing that I did. really appreciate your reply, will definite show off the loop once I get to do it!
Wow nice to see the thermal performance! That's what I will be challanging with my 240+280 negative pressure plan. The build is alomst there. I'm going to build it after I take my GRE test on 10.16, just a few days left!
Could you provide the CPU + GPU power corresponding to those tests? That will provide a better reference to decide which is better.
BTW the 240 + 280 negative pressure plan may not be the end! I just found a possible way to actively cool a 280*30mm radiator at GPU side. I need to wait until I get the case to find out whether I can really make it. Hoping that to enable us to extremely overclock everything in this 9.5L case (maybe 250W CPU + 400W GPU)!
Similar threads
