Production Aquanaut Basic/Extreme - Ultra Low Profile CPU Block & Pump Mount Combo

duynguyenle

Airflow Optimizer
Aug 20, 2019
331
331
Uhh the screw in that location actually reduces stress in the acrylic mid span. So no, adding a hole there wont increase the chances of a crack occurring. Pyotor has a valid point, in that the bow in the acrylic loads the sealing ring unevenly. Orings are recommended to have a optimal squish for sealing against set pressures. Loading unevenly means some areas are over squished, some is optimal and some is under squished. Also acrylic is very good at creep so yes there is a possibility of leakage at a later date due to deformation of the plastic and loss of preload in the center. Creep occurs at room temperature for loaded plastics, and heat accelerates the process.

I dunno, based on that photo, the pump attachment bolt right above it, so you'd be loading the region between those two holes in shear. There's too many variables in this particular example (especially regarding the flatness of the pump housing itself) so I can't really say with any certainty whether this additional shear in the corner is of concern.

In terms of creep, from my (very limited) experience with plastics (mainly non-structural POM and PEEK), I would be more concerned about thermally induced creep at regions of higher stresses (fastener holes/notches, and especially holes with threadforms, in this case). You are correct that on a longer timeframe, loss of preload due to creep could be a problem, I'm just not entirely convinced that it is of the same order of importance as the loss of preload due to thermally induced creep around threadforms ( I don't have as much experience with acrylics compared to POM/PEEK/Nylon).

In any case, I stand by my point about the modded hole. One way to improve that would be to actually counterbore the bottom plate and use a bolted joint with a nylon washer instead of tapping directly into acrylic (but I am not convinced there's adequate margin in the bottom plate to do so)
 

Pyotr

Caliper Novice
Oct 2, 2020
22
9
Ouch, that looks like a crack waiting to happen. Drilling and tapping a hole with just a few mm of acrylic on either side in the area of the block with the most stress? That doesn't sound like a good idea to me, especially when the screw will just be holding against threads cut into the acrylic. I would rather deal with changing an O-ring every few years (the loop should be torn down for cleaning anyhow) than ruining a perfectly good block.
True. I did state its a less than idea location to install. The chamfer from the head does reduce, but not eliminate stress risers in the material. over time with heat cycles it drastically increases the risk. It's why I'd rather look into extending the coldplate screws into the top plate. Alternatively, using a material that's not as susceptible to cracks E.G. POM. (have I beaten that horse to death yet?)

Uhh the screw in that location actually reduces stress in the acrylic mid span. So no, adding a hole there wont increase the chances of a crack occurring. Pyotor has a valid point, in that the bow in the acrylic loads the sealing ring unevenly. Orings are recommended to have a optimal squish for sealing against set pressures. Loading unevenly means some areas are over squished, some is optimal and some is under squished. Also acrylic is very good at creep so yes there is a possibility of leakage at a later date due to deformation of the plastic and loss of preload in the center. Creep occurs at room temperature for loaded plastics, and heat accelerates the process.
Exactly. the trade off with the location is risk cracking the acrylic, or risk blowout of the seals. given that you can see the consequence of the bend in the sealing profile of the preformed packing, id rather risk potential cracks than a blowout.


 

Nouvolo

Creator
Original poster
Sep 8, 2018
759
1,712
www.nouvolo.com
The mod looks ok I think.

Aquanaut was designed with all screws installed from the bottom for better aesthetics in mind.

Mentioning again, there is no leak issue with this, after extensive testing. Feel free to do whatever for your peace of mind😉. Just don't want after about 50 posts down the thread, users will get confused and everybody starts trying to mod, thinking the original will leak...
 

einmannbude

Caliper Novice
May 22, 2016
25
18
Yesterday I finished assembling my custom loop with Aquanaut as CPU block. Unfortunately the CPU temperatures are beyond good and evil. I hope, I made some installation error, but just to make sure this is not expected behaviour, I wanted to share the data I get and ask for feedback:

In Windows idle the CPU is already nearly 60°C hot, if I let pump and fans run at reasonable speed for idle mode.
If I max out fans and pump, the water temperature is around 27°C in Windows idle. CPU temps around 50°C.

So, there is a delta between CPU and water of more than 20°C. This should not be the case, right? At the same time my GPU is at 30°C in idle, so there is almost no delta.
 

einmannbude

Caliper Novice
May 22, 2016
25
18
Yesterday I finished assembling my custom loop with Aquanaut as CPU block. Unfortunately the CPU temperatures are beyond good and evil. I hope, I made some installation error, but just to make sure this is not expected behaviour, I wanted to share the data I get and ask for feedback:

In Windows idle the CPU is already nearly 60°C hot, if I let pump and fans run at reasonable speed for idle mode.
If I max out fans and pump, the water temperature is around 27°C in Windows idle. CPU temps around 50°C.

So, there is a delta between CPU and water of more than 20°C. This should not be the case, right? At the same time my GPU is at 30°C in idle, so there is almost no delta.
Nevermind, I found the problem. It was wrong installation indeed. The CPU block didn't even touch the CPU. Now everything works very well.
 
  • Like
Reactions: zacobin and rfarmer

sos

Chassis Packer
Feb 11, 2021
15
20
@Nouvolo Is there a possibility to rotate am4 bracket by 90 degree?




Or they may be a problem with tightening socket screws?
 

Valantar

Shrink Ray Wielder
Jan 20, 2018
2,201
2,225
Finally got my pump (Laing DDC 3.2 PWM) wired up + sleeved, and got the Aquanaut assembled today. Fitment looks excellent on my part, no gaps, bends or flexing visible. Very happy so far, looking forward to getting it installed.

Also decided to sand the surface of its screws for some extra grip + a less shiny look. About 10 minutes and a sheet of 240 grit sand paper and IMO they look pretty decent. Not professional-level obviously, but suitably matte, and definitely grippier. The camera's sharpening does exaggerate how scratchy they look, as the real-world finish is more of a vague criss-cross scratch pattern in a mostly matte surface. It's worth noting that sanding brings out any and all imperfections very clearly though - the flatter part on the top of the screw in particular. But I like it!
 

Nouvolo

Creator
Original poster
Sep 8, 2018
759
1,712
www.nouvolo.com
Apr 2021 Update:

We now accept different currencies from major countries/regions, including
🇺🇸
USD,
🇪🇺
EUR,
🇯🇵
JPY,
🇬🇧
GBP,
🇨🇦
CAN,
🇦🇺
AUD,
🇭🇰
HKD & more.

And we now have distribution for EU countries with prices include VAT. This excludes UK due to brexit, UK buyers will need to pay VAT themselves when taking delivery. However we are working on UK VAT registration.

Find out more: https://www.nouvolo.com
 

Valantar

Shrink Ray Wielder
Jan 20, 2018
2,201
2,225
Guess I can share some of my experiences from running the Aquanaut on my 5800X.

First off, just to be clear, the 5800X is a pretty hot running CPU - being a single CCD 105W TDP / ~140W boost CPU it has far, far higher thermal density than the other Zen3 CPUs, both the 65/78W 5600X as well as the dual-CCD 105/144W 5900X and 5950X. That makes it by far the most difficult Zen3 CPU to keep cool. I run mine completely stock in an ASRock B550 Phantom Gaming ITX.

Also, I don't really have any useful comparisons to make. Now that I have QDCs in the loop I might swap my old EK block back in at some point (and just use the Aquanaut as a pump for that testing), but that's far too much work for me to realistically do in the near future. Comparisons are also made extremely difficult through the heaps of other variables in play - a GPU in the loop, the amount of radiators, fans and fan speeds, not to mention the fluctuating thermal loads inherent to contemporary CPUs and their dynamic boost algorithms. I also have my setup tuned for silence most of all, which necessitates allowing for higher temperatures when there's a 275W GPU in the loop and only a single 240mm radiator with 1400rpm max fans. The PWM response of my DDC pump is also rather unpredictable (the same PWM number can result in 2-300rpm difference depending on whether it's slowing down to that number or speeding up to it). My fans and pump speed are controlled by water temperature in order to avoid rapid fan/pump ramps, so CPU thermals are by design allowed to spike quite significantly in my setup.

With the caveats out of the way, how does the Aquanaut fare? I guess I would say "fine". To expand a bit on that:
- My desktop usage temps, with the pump at ~1200rpm and just a single fan on the rad running at ~800rpm, sit in the mid 40s to low 50s. Spikes from opening an application or something similar can easily send them into the mid 60s, before settling down again. Note: this is a very low pump speed. There are very noticeable temperature improvements from increasing pump speed to 2000-2500rpm.
- While gaming, the whole loop heats up a lot more, so the pump and fans obviously run faster, and all fans in the system (two on the rad, two exhaust in the case) are on and running, and the pump ramps first to ~1800rpm and then ~2500rpm depending on thermals (there are some weird resonances at in-between speed, hence the stepped approach). CPU temperatures when gaming are typically in the 60s, but I've seen peaks around 80. This is of course hugely dependent on the game in question, and is also massively affected by the GPU pumping 275W of heat into the loop, significantly raising the base temperature possible for the CPU. My radiator fans don't ramp to full speed until the coolant after the radiator reaches 40°C, which it pretty much never does, so they're typically running at 1100-1200rpm while gaming.
- CPU-intensive loads like Lightroom exports can make thermals spike pretty high, and I've seen it pass the 80°C under these conditions. But again, due to the fans and pump being controlled solely through liquid temperature, it's slow to compensate for this by design. There's no doubt it could handle these spikes better with slightly higher pump and fan speeds.

Here's a screenshot of Aquasuite while I'm writing this post, on the desktop, with a ton of Firefox windows open, a few game launchers and other stuff in the background, as well as a few Word documents and Outlook - that's about as "idle" as this PC gets.


This is a pretty specific setup, so it's probably not very applicable to people's use cases, but the main takeaway is. It's pretty simple, really: don't expect miracles. This is a competent cooler that does its job decently and provides compact mounting for a high flow rate pump - my GPU temperatures have actually dropped noticeably with this setup, despite the previous setup having a less power hungry CPU (1600X) and an additional 120mm rad that I could no longer fit. It doesn't run the CPU at crazy low temperatures under high loads, but it can allow for impressively silent setups in compact builds if that's what you're after.

Edit: spelling
 
Last edited:

ezerez

Trash Compacter
Mar 31, 2021
40
34
Guess I can share some of my experiences from running the Aquanaut on my 5800X.

First off, just to be clear, the 5800X is a pretty hot running CPU - being a single CCD 105W TDP / ~140W boost CPU it has far, far higher thermal density than the other Zen3 CPUs, both the 65/78W 5600X as well as the dual-CCD 105/144W 5900X and 5950X. That makes it by far the most difficult Zen3 CPU to keep cool. I run mine completely stock in an ASRock B550 Phantom Gaming ITX.

Also, I don't really have any useful comparisons to make. Now that I have QDCs in the loop I might swap my old EK block back in at some point (and just use the Aquanaut as a pump for that testing), but that's far too much work for me to realistically do in the near future. Comparisons are also made extremely difficult through the heaps of other variables in play - a GPU in the loop, the amount of radiators, fans and fan speeds, not to mention the fluctuating thermal loads inherent to contemporary CPUs and their dynamic boost algorithms. I also have my setup tuned for silence most of all, which necessitates allowing for higher temperatures when there's a 275W GPU in the loop and only a single 240mm radiator with 1400rpm max fans. The PWM response of my DDC fan is also rather unpredictable (the same PWM number can result in 2-300rpm difference depending on whether it's slowing down to that number or speeding up to it). My fans and pump speed are controlled by water temperature in order to avoid rapid fan/punp ramps, so CPU thermals are by design allowed to spike quite significantly in my setup.

With the caveats out of the way, how does the Aquanaut fare? I guess I would say "fine". To expand a bit on that:
- My desktop usage temps, with the pump at ~1200rpm and just a single fan on the rad running at ~800rpm, sit in the mid 40s to low 50s. Spikes from opening an application or something similar can easily send them into the mid 60s, before settling down again. Note: this is a very low pump speed. There are very noticeable temperature improvements from increasing pump speed to 2000-2500rpm.
- While gaming, the whole loop heats up a lot more, so the pump and fans obviously run faster, and all fans in the system (two on the rad, two exhaust in the case) are on and running, and the pump ramps first to ~1800rpm and then ~2500rpm depending on thermals (there are some weird resonances at in-between speed, hence the stepped approach). CPU temperatures when gaming are typically in the 60s, but I've seen peaks around 80. This is of course hugely dependent on the game in question, and is also massively affected by the GPU pumping 275W of heat into the loop, significantly raising the base temperature possible for the CPU. My radiator fans don't ramp to full speed until the coolant after the radiator reaches 40°C, which it pretty much never does, so they're typically running at 1100-1200rpm while gaming.
- CPU-intensive loads like Lightroom exports can make thermals spike pretty high, and I've seen it pass the 80°C under these conditions. But again, due to the fans and pump being controlled solely through liquid temperature, it's slow to compensate for this by design. There's no doubt it could handle these spikes better with slightly higher pump and fan speeds.

Here's a screenshot of Aquasuite while I'm writing this post, on the desktop, with a ton of Firefox windows open, a few game launchers and other stuff in the background, as well as a few Word documents and Outlook - that's about as "idle" as this PC gets.


This is a pretty specific setup, so it's probably not very applicable to people's use cases, but the main takeaway is. It's pretty simple, really: don't expect miracles. This is a competent cooler that does its job decently and provides compact mounting for a high flow rate pump - my GPU temperatures have actually dropped noticeably with this setup, despite the previous setup having a less power hungry CPU (1600X) and an additional 120mm rad that I could no longer fit. It doesn't run the CPU at crazy low temperatures under high loads, but it can allow for impressively silen setups in compact builds if that's what you're after.
Thx for the write up. Wouldnt call your setup that specific though. Powerfull workstation in a small form factor that runs silent in most conditions and can ramp up but keep acceptable noise and thermals when gaming is exactly what I am also planning to build. Whats the GPU thats putting 275W in the loop? Im gonna probably build a 5600x with a 3080 combination. So if your loop handles that then it shouldnt be a problem with a 5600x. What radiator are you using and which DDC pump did you put on the aquanaut if I may ask?

kind regards
 
  • Like
Reactions: sos

Valantar

Shrink Ray Wielder
Jan 20, 2018
2,201
2,225
Thx for the write up. Wouldnt call your setup that specific though. Powerfull workstation in a small form factor that runs silent in most conditions and can ramp up but keep acceptable noise and thermals when gaming is exactly what I am also planning to build. Whats the GPU thats putting 275W in the loop? Im gonna probably build a 5600x with a 3080 combination. So if your loop handles that then it shouldnt be a problem with a 5600x. What radiator are you using and which DDC pump did you put on the aquanaut if I may ask?

kind regards
Thanks for the feedback :)

  • The pump is a Laing DDC 1T PWM, which from what I can tell is the same as the DDC 3.2 PWM. I've stuck an Alphacool heatsink on it, and re-crimped the wiring for it to be powered off one of the headers on my Aquacomputer Quadro. The Quadro is btw an excellent investment for good fan control, especially with its thermal sensor inputs.
  • The GPU is my old (old!) Radeon Fury X, with an EK block on it. I've got a 6800 XT on order, but who knows when that will show up, and I don't think there are any waterblocks available for that model yet anyhow. But it's definitely reached retirement age :p
  • The radiator is an EK PE (38mm) 240mm, and the fans are Be Quiet Silent Wings 3s (not the high speed version). I'm planning on moving to a 280mm rad with (probably) Arctic P14 fans whenever the Meshlicious becomes available in Europe though.
 
  • Like
Reactions: sos

ThomasThailand

Efficiency Noob
Apr 30, 2021
5
13
@ThomasThailand Why you want to swap Ryzen to 5600x? 🤔 And what about fans rpm at full load?
It s true that it s too noisy at full rpm and the 5950x suppose to go in my o11d xl with 3 rads. With the 5600x, it s gonna be much more silent.
I didn't precise that temp are okay for gaming/work renders but it throttle in long benchmark especially when avx is activated. I just wanted to know the max capacity of this system with such a big processor, itx and rad size are much adapts to 5600x or 5800x max.
 

aquelito

King of Cable Management
Piccolo PC
Feb 16, 2016
952
1,124
Thanks for the feedback :)

  • The pump is a Laing DDC 1T PWM, which from what I can tell is the same as the DDC 3.2 PWM. I've stuck an Alphacool heatsink on it, and re-crimped the wiring for it to be powered off one of the headers on my Aquacomputer Quadro. The Quadro is btw an excellent investment for good fan control, especially with its thermal sensor inputs.
  • The GPU is my old (old!) Radeon Fury X, with an EK block on it. I've got a 6800 XT on order, but who knows when that will show up, and I don't think there are any waterblocks available for that model yet anyhow. But it's definitely reached retirement age :p
  • The radiator is an EK PE (38mm) 240mm, and the fans are Be Quiet Silent Wings 3s (not the high speed version). I'm planning on moving to a 280mm rad with (probably) Arctic P14 fans whenever the Meshlicious becomes available in Europe though.

Thanks a lot for the detailed feedback :)

This is indeed a very low pump speed ; Martin's lab has showed that performance really start to deteriorate below 0.4 GPM.
Have you been able to monitor the flow ?

I'm also pretty picky about noise and to me, the DDC 3.2 PWM only starts to become annoying above 35% power.
Maybe the fact that it is hard mounted onto the waterblock, that is to your case, makes things worse noise-wise ?

Anyway, I was planning to update my Lian Li PC-TU100 with the Aquanaut and a dual rad setup : one rear 80mm rad for the GPU (GTX 1650 Super), one front 120mm rad for the CPU (3700X).
The 80mm rad wad modded to add a fill port on top. No need for a res :)

Any chances to get the 3D file of the Aquanaut @Nouvolo ?
 

Valantar

Shrink Ray Wielder
Jan 20, 2018
2,201
2,225
Thanks a lot for the detailed feedback :)

This is indeed a very low pump speed ; Martin's lab has showed that performance really start to deteriorate below 0.4 GPM.
Have you been able to monitor the flow ?

I'm also pretty picky about noise and to me, the DDC 3.2 PWM only starts to become annoying above 35% power.
Maybe the fact that it is hard mounted onto the waterblock, that is to your case, makes things worse noise-wise ?

Anyway, I was planning to update my Lian Li PC-TU100 with the Aquanaut and a dual rad setup : one rear 80mm rad for the GPU (GTX 1650 Super), one front 120mm rad for the CPU (3700X).
The 80mm rad wad modded to add a fill port on top. No need for a res :)

Any chances to get the 3D file of the Aquanaut @Nouvolo ?
I think it's likely that the mounting is affecting noise levels, as it's pretty evident that at least some of the noise is resonant noise rather than pure motor noise, and appears and disappears at various speeds. 1200-1400rpm is dead silent, 30%/1900rpm is really annoying (a kind of ringing, resonant hum that feels like it's coming from inside of my head), but that mostly goes away again around ~34%/2300rpm. I could probably run 35% at idle, but since the pump is audible (even if rather inoffensive) at that speed and I'm pretty sensitive to those types of noises I prefer to run it lower at low loads. I have no idea what the flow rate is at these speeds, but it is at least sufficient to keep the system working fine, and of course it ramps up once there is enough load on the system to heat up the fluid. Motor noise starts getting annoying again at 42-43%/3300rpm, though there's not the same resonant hum at those speeds. 42-43% is still more annoying than 55%/4400rpm full speed though. It's all pretty weird, so my profile is tuned to avoid these noisy ranges. I'll no doubt keep tuning this over time, but so far I'm pretty happy with my current profile. We'll see what changes need to happen when I move to the Meshlicious, if nothing else :)
 
  • Like
Reactions: aquelito

NGOD

Minimal Tinkerer
New User
Aug 7, 2019
3
0
@Nouvolo Hi, is there any chance that you can ship this to Argentina? I tried to contact a few forwarding service to get this but they want to charge me a stupid amount of money. Thanks