Production Aquanaut - Ultra Low Profile CPU/Pump Water Block

Nouvolo

Creator
Original poster
Silver Supporter
Sep 8, 2018
508
1,020
www.nouvolo.com
Someone can confirm if I got a measurement wrong, but IIRC 41mm is lowest possible:
  • Aquanaut itself is 21mm tall
  • Pump, without top, is 20mm tall -> absolutely no heatsink and just the thin plastic cover
  • Fittings installed height less than 20mm tall -> the only fittings I know are Koolance Elbow rotary and Alphacool TPV 90 degree rotary
Aquanaut height is about 24mm including the cold plate. The acrylic blocks are about 21mm tall.
 
  • Like
Reactions: thelaughingman

cennis

Trash Compacter
Jan 13, 2020
37
7
Someone can confirm if I got a measurement wrong, but IIRC 41mm 44mm is lowest possible:
  • Aquanaut itself is 21mm 24mm tall
  • Pump, without top, is 20mm tall -> absolutely no heatsink and just the thin plastic cover
  • Fittings installed height less than 20mm tall -> the only fittings I know are Koolance Elbow rotary and Alphacool TPV 90 degree rotary
EDIT: revised to 44mm as I missed the 3mm coldplate
Bykski 90 degree non rotary are 18.5mm, each one is threaded differently so if you try enough of them it should point the direction you want.
 
  • Like
Reactions: thelaughingman

duynguyenle

Airflow Optimizer
Aug 20, 2019
279
256
Bykski 90 degree non rotary are 18.5mm, each one is threaded differently so if you try enough of them it should point the direction you want.

I have mistakenly bought Bykski non-rotary 90deg fittings in the past. Do not recommend, especially if you're screwing them into POM/Acrylic, it's nigh impossible to have them tightened properly and still point in the direction you want, due to the fixed non-rotary nature of the fitting. Buyers beware.
 

Poblopuablo

King of Cable Management
Bronze Supporter
Jan 14, 2018
794
444
Dang. :(
Someone can confirm if I got a measurement wrong, but IIRC 41mm 44mm is lowest possible:
  • Aquanaut itself is 21mm 24mm tall
  • Pump, without top, is 20mm tall -> absolutely no heatsink and just the thin plastic cover
  • Fittings installed height less than 20mm tall -> the only fittings I know are Koolance Elbow rotary and Alphacool TPV 90 degree rotary
EDIT: revised to 44mm as I missed the 3mm coldplate
 

ToWi1989

Chassis Packer
Aug 26, 2020
14
16
Since im still waiting for the ekwb 3090 fe block i decided to go ahead and mock build the pc anyways. Im using the aquanaut with a 5950x but im not too impressed. First of all some of the fins of the cold plate are not machined completely, affecting the contact surface area with the coolant. The top plexi plate arrived full of small hair scratches, so it wouldve been better off being matte like the lower plate, or, even better, in black acetal.

The temps are not great but not bad. On 1t load with stock overboost, the temp peaks to 67C and theres quite a big delta between package temp and cpu temp. Manual all core OC seems to be quite decent, but PBO uses a much higher voltage to reach the same clocks as a manual OC, leading to much higher temps. This is using 2x 240mm st30 rad, without the 3090 in the watercooling loop! Still contemplating if itll be worth producing my own block (non reversed flow) with a decent cold plate.
 

cm33ks

Case Bender
New User
Dec 5, 2020
2
0
I can echo the disappointment with the quality of the fins as well as the scratches in the acrylic. Why not design a pump top for an existing EK block?
 

Nouvolo

Creator
Original poster
Silver Supporter
Sep 8, 2018
508
1,020
www.nouvolo.com
Since im still waiting for the ekwb 3090 fe block i decided to go ahead and mock build the pc anyways. Im using the aquanaut with a 5950x but im not too impressed. First of all some of the fins of the cold plate are not machined completely, affecting the contact surface area with the coolant. The top plexi plate arrived full of small hair scratches, so it wouldve been better off being matte like the lower plate, or, even better, in black acetal.

The temps are not great but not bad. On 1t load with stock overboost, the temp peaks to 67C and theres quite a big delta between package temp and cpu temp. Manual all core OC seems to be quite decent, but PBO uses a much higher voltage to reach the same clocks as a manual OC, leading to much higher temps. This is using 2x 240mm st30 rad, without the 3090 in the watercooling loop! Still contemplating if itll be worth producing my own block (non reversed flow) with a decent cold plate.
Please check out optimum tech's review. Ali's temp is near 70°C with CPU synthetic load using standard custom loop config...(Alphacool Eisbaer LT) I don't know what to say...


Are the rads installed in T1? Or is it open bench? Appreciate If you could use the same setup to compare with other water blocks. IF rads are in T1, I suspect air flow is more of a determining factor.

There are earlier posts on more detailed testing with Aquanaut, please refer to those...
 
Last edited:
  • Like
Reactions: thelaughingman

ToWi1989

Chassis Packer
Aug 26, 2020
14
16
Im not sure why you are refering to this vid. My build specs are as followed:

Case: Ghost S1 with large top and bottom hat
Mobo: Asus ROG STRIX X570-I
RAM: G-Skill F4-4000C18-64GTZR
CPU: AMD 5950X
GPU: Nvidia 3090 FE
PSU: Corsair SF750

4x noctua A12x25 PWM fans
2x alphacool 240mm ST30 rad
2x meta customes mount plates for the S1
DDC 3.2 pump
And obviously the aquanaut

Custom wiring to avoid air flow restrictions as much as possible

Im currently running prety much an open test bench. The rads are mounted to the S1 skeleton, but the top and bot hats and side panels are not mounted. The bottom rad is generously spaced from the desk surface.

This morning I refitted the block and checked thermal paste, but all looked good (this is a PITA to keep having to do since im using hardtubing and dont have any soft tube fittings laying around). I've tried a couple of benchmarks. With PBO set to PPT 180W, TDC 160A, EDC 190A and curve optimizer to -10 on all cores, during a cinebench R23 run the temp spikes to 80-83 deg C and stays there (~4.3 ghz all core). I've played around with my fan settings, and I'm able to decrease the fan speed to around ~40% pwm before temps start to creep up. This leads me to believe that the heat transfer from the cold plate is bottlenecking the cooling performance. Looking at other cold plates (apogee drive / EKWB) they have a MUCH bigger surface area.

I dont have any other blocks laying around for testing.
 

Nouvolo

Creator
Original poster
Silver Supporter
Sep 8, 2018
508
1,020
www.nouvolo.com
Im not sure why you are refering to this vid. My build specs are as followed:

Case: Ghost S1 with large top and bottom hat
Mobo: Asus ROG STRIX X570-I
RAM: G-Skill F4-4000C18-64GTZR
CPU: AMD 5950X
GPU: Nvidia 3090 FE
PSU: Corsair SF750

4x noctua A12x25 PWM fans
2x alphacool 240mm ST30 rad
2x meta customes mount plates for the S1
DDC 3.2 pump
And obviously the aquanaut
...
- Try under volt for better temps. PBO (auto mode) is known for generating tons of heat. Personally I have tried PBO I don't trust it to be that power efficient for any given clock
- Check for bubbles in the loop, especially within the block, you may need to move your rig about and adjust pump speed to guide bubbles out of the loop. Make sure there is no bubbly, sloshy sound

For 5950X, if you can get temps down to about 70°C, I think it is a decent temp. Maybe more tweaking is required for such top end extreme setup.

Another video from Ali with temp issues for 5950X & RTX 3090. Understand the case used and config are different. But seems the CPU/GPU combo is generating lots of heat.
 
Last edited:

ToWi1989

Chassis Packer
Aug 26, 2020
14
16
Ooo yes. I agree that PBO doesnt yield the best results looking at clocks vs heat, but its the only way to combine 1T and all core OC without having to manually switch profiles all the time depending on your workload. Also, with the power limit set in PBO you can control the heat generated by the CPU prety well. For example; running prime95 (small FTT) generates an absolute shitload of heat. Compared to a stable cinebench run, i need to up the voltage from 1,05 to 1.125V (using manual OC), but with PBO and a fixed power limit it just reduces the clock speed. At this moment I'm running with a 160W limit and that seems to give OK temps/clocks (65C during gaming and ~4.2 ghz all core in cinebench).

Bubbles are almost completely elimitated. I use a vacuum pump and a long tube connected to one of the ports on the rad and 'guide' all air pockets to the outlet by rotating/shaking the complete pc. There's still a very small 'dead' pocket, but no circulating air thanks to a low pump speed (20-25% pwm).

Given that I can run very low fan rpm with just the CPU in the loop gives me some hope adding the 3090 will not have a huge impact on cpu temps, as long as the additional heat generated can be dissipated by upping the fan rpm to keep water temp under control. I have a temp sensor in backorder to be able to monitor water temps and controlling fan speed.

I never thought Ali would be able to have a working setup with just 1 rad, so i wasnt surprised with the temps he is seeing. Thats a whole different bottleneck limiting his system :)

Anyways, all I want to say is that I believe there's more performance to be had with a better cold plate :)
 
Last edited:

ToWi1989

Chassis Packer
Aug 26, 2020
14
16
  • Like
Reactions: Nouvolo and rfarmer

Lbibass

Chassis Packer
May 2, 2020
16
7
I currently have my aquanaut in my FormD T1, and while I quite like it, temps seem quite a bit higher than I would have expected, especially given that I am running a 240mm radiator. Generally, in aida64, I get temps of around 90-95 degrees on my 3900x at stock settings. I would have expected better from a DDC and a large radiator. Oneof the more concerning things is that the difference in temps between the CCDs on said 3900x are generally in excess of 10 degrees, which seems quite significant. It's almost as if the coldplate isn't very level. Has anyone else had a similar experience?
 

Nouvolo

Creator
Original poster
Silver Supporter
Sep 8, 2018
508
1,020
www.nouvolo.com
I currently have my aquanaut in my FormD T1, and while I quite like it, temps seem quite a bit higher than I would have expected, especially given that I am running a 240mm radiator. Generally, in aida64, I get temps of around 90-95 degrees on my 3900x at stock settings. I would have expected better from a DDC and a large radiator. Oneof the more concerning things is that the difference in temps between the CCDs on said 3900x are generally in excess of 10 degrees, which seems quite significant. It's almost as if the coldplate isn't very level. Has anyone else had a similar experience?
best way to check the cold plate is just to take it off. If the thermal paste pattern is spread out around the CPU then it should be fine. 😉 Bottomline, you just want contact.
 

NRG

Cable-Tie Ninja
Oct 30, 2015
213
245
@Nouvolo

Hi. Early buyer wondering how i go about getting the screw head hardware as the original ones are frustrating. Thanks
 

evd

Trash Compacter
Mar 24, 2020
37
13
Got my Aquanaut delivered today to Australia! looking forward to seeing how it goes in my loop to replace the LT Solo - THat thing pumps like a trickle and makes filling PAINFUL.

Hoping this will make it a lot easier and run it pretty slow for some noise free operation...Will check out the LED mods that people have done previously as well..

Just need my Alphacool DDC to arrive.
 
  • Like
Reactions: Nouvolo

Torquednb

Efficiency Noob
Nov 16, 2020
5
24
I mounted the Aquanaut this weekend and did a pressure test using the Dr. Drop, there was a small leak at the inlet o-ring. The image shows the Aquanaut leak free but i'm concerned about the acrylic bowing.
The 2 acrylic parts aren't touching so that should be oke, i think a POM or extra screws would be nice 😉

Despite my concerns i still think this is a great product for the SSF community
 
  • Like
Reactions: Shinken1