Dschijn

A friend replaced all his Noctua NF-A14 with the new Arctic P14. he got a quieter setup that cools better. These fans are great.

The new Arctic P12 PWM PST in black? They are really good! Cheap, very quiet, high performance.

The new Arctic P12 PWM PST in black? They are really good! Cheap, very quiet, high performance.

It's ~1°C which is not much. With high pump speeds it might be even less.

Yes will be enough, but not super cool nor super quiet.

He needs this:
 
 

He needs this:
 
 

So just over 3 years ago, in this topic, I decided to take a stab at doing my own iteration of the "Whole Room Water Cooling" project that LTT did just over 3 years ago. I have a small office that heated up fairly quickly with my rig expelling a bunch of heat into the room. My "Basement Water Cooling Project" actually didn't fix that problem. Yes, the office does heat up a little slower, but 3 monitors output a surprising amount of heat. But heat wasn't the sole reason I did this project. The main reason I did this project was to rid my office of fan noise, as I do video and audio work for a living, and I wanted my office to be as silent as possible. And in that regard, the project was a massive success. But three years have gone by, so I wanted to make a new topic about it, rather than bumping a thread that hasn't had a reply for over 2 years.

So, there's really only 2 questions that I feel need answering.

1. Is the "Basement Water Cooling Project" still working?
2. Was it worth it?

So, is the project still working? Short answer, yes. Long answer, yes, and it has required shockingly little maintenance. Since I properly completed the project about 2 1/2 years ago, I've only recycled the water once, which was in November of 2017, when I upgraded my rig. But even then, and now, the water inside of the system has remained almost completely clear, but there is a sort-of new exception.



Judging by this image, which I took under an hour ago, things look great! At a glance, the water looks completely clear. But if we take a closer look at the outlet from the pump, as well as the inlet for the reservoir, something looks off.



If you know anything about clear tubing, it shouldn't be too hard to see that the tubing shouldn't have a blue tint when there's a brown desk right underneath it. So something is going on inside of the system. But what?



Well I think this image sheds a lot of light on what's going on. The acrylic reservoir shows no signs of tinting at all, leading me to believe that the biocide that I used (PT_Nuke), which can be seen in this image, has begun to stain the tubing used in the rest of the loop. So, am I going to do anything about this? Honestly, probably not. There's no hint of any corrosion of any kind, and the water is pretty much crystal clear. So, 3 years, one flushing of the system without ever replacing the tubing.

But was it worth it?

Well considering that I've lived in this house for over 3 years now, and don't plan on moving any time soon, I would say yes. However, I spent a LOT of unnecessary money on this project, primarily when I bought my original pump, which it, along with all of my other mistakes, can be found in the original thread.

I'm a big ol' geek. I found this project to be a lot of fun, and I'd do it again. I doubt LinusTechTips will ever re-approach their project, but if they did, I can confidently say that as long as they use spring water, a biocide like PT_Nuke, and they don't mix in an aluminum reservoir or any other aluminum, and they stick to copper / nickel (and I think brass is fine too, as I think my fittings on the copper pipe are brass), then they'd have a system that would require a shockingly low amount of maintenance.

Anywho, was randomly in the mood to post an update, so I thought I'd do so. Thanks!

(The computer's lights go through different shades of fairly deep blue colors, but always show up purple in images, for whatever reason)

Among the best: EK-CoolStream XE 480

largely depends on the coolant you use. Every year would be very safe for rhe modern non-opaque coolant like EK cryofuel, they can probably even go longer.

EK SE rads are not the best rads... If it has to be EK, get the PE ones.

Keep in mind that the QUADRO only can handle PWM fans/pumps!

He means that he went through other liquids in the same loop without having this problem so any mechanical cause is unlikely.

Please read before commenting.

Small update:
I decided to change the front and top panel with a custom cut perforated aluminum mesh.
Painted in white or entirely black, maybe with some Lian Li logo, have to see.
That should solve the problem for a good amount hopefully.
 
 
 

I switched the top fans back to intakes and it helped by a couple degrees, but I'm still not satisfied. I'd like to try all the radiator fans as exhaust, but that would require taking the whole thing apart, due to the lowest radiator fan being inaccessible. Might just have to go with a different case.
 
Trying to balance noise, price, looks, size, and performance all at the same time really is difficult.

Guys... that are the exact same fans. Just the rpm range is different. So at the same rpm they will perform the same . You just have to get an idea which rpm range is the one that suits you and will match the rpm you want to stay at.
If you are open for other fan suggestions: Noctua NF-A12x25 (much quieter than the SW3)

The only way to keep the room cooler would be to promote airflow through it to remove the heat. Watercooling makes it more efficient at keeping the components cooler but will not cool down or heat up the room more since your heat output from the components is still the same. 
 

Yayy, test time. How much does the airflow through a radiator really matter??
 
I got 4x PFC1212DE-PWM high speed "server" fans for fun, decided to try them on my 480 mm radiator and do a comparison against regular Noctuas on a 360 mm.
 
Test setup:
Procedure
- Main heat load produced with CPU running Prime95 small FFT AVX
- Tested both idle and load, with only 3x Noctuas and with only 4x Deltas - other radiator fans off, pumps max.
- Monitored: ambient temp (in the video both in and out are in the room), water temp (inline sensor), CPU cores temp, PSU power output (Delta fans ran from a different PSU and were not included in the measurement).
 
System
CPU: i7 3970X, load @ 4.7 GHz, 1.40 V
GPU: HD 7850 idling
PSU: Corsair AX860i (core temps & power monitoring 1/sec)
 
Water loop
Rad1: Alphacool UT60 360 mm (with 3x Noctua)
Rad2: Alphacool UT60 480 mm (with 4x Delta)
Rad3: Phobya Xtreme Nova 1080 mm (with 9x Corsair ML, off all the time)
Pumps: 3x Laing D5 vario in series
1x Alphacool CPU block
1x Alphacool universal block
 
Fans
- 3x Noctuas (2x NF-F12 PWM + 1x  NF-P12 PWM)
- 4x Delta PFC1212DE-PWM

 
Note: Deltas were mounted with shrouds made diy-style from some 120 mm fan rims in hopes of reducing the large dead spot in the center.
 
All runs (1-4. 1st day, 5-7. 2nd day)
1. Idle - Noctuas idle, Deltas off, pumps min
2. Idle - Noctuas idle, Deltas off, pumps max
3. Idle - Noctuas max, Deltas off, pumps max
4. Idle - Noctuas off, Deltas max, pumps max
5. Idle - Noctuas max, Deltas off, pumps max
6. Load - Noctuas max, Deltas off, pumps max
7. Load - Noctuas off, Deltas max, pumps max
 
Note: During run 6. water temp rising from idle, during run 7. water temp cooling from run 6 end. Ambient slowly rising as the room warms up (no A/C).
 
Results

 
Note:  Ambient and water temp taken at the end of the runs. CPU cores temp is all cores average taken over a small time period at the end of run. Power is the whole average output during the run. Idle includes the open programs and logging. Accuracy is what it is. At small differences accuracy suffers, but gives the overall idea. Rad size compensated values calculated by increasing the ambient-water ∆ proportionally to rad size difference (marked with red).
 
Video runs 5-7
 
Some observations
- At load Deltas lowered water ∆ from 11.5C to 4.3C when comparing rad size compensated values (167% more powerful cooling)
- Changing pumps from setting 1 (min) to 5 (max) increased power draw by ~40 W
- Idle to full CPU load increased the system power draw by about 320 W.
- When on, the Delta fans themselves should warm the room air with about 200 W of heat output
 
Bonus clip for whoever made it down here
 
Happy water cooling!
 
I will be running with the Deltas at full blast from now on.

They deliver it with 2 different length of screws. If you need other screws get M4 threaded ones.

Some Poseidon cards are cooling very badly!

Delid if you want to OC and get close to 5GHz.

Pretty sure you want the pump at the lowest point preferably to eliminate any issues with air bubbles or something like that.

Overall doesn't matter too much.

Figure you guys might enjoy some custom waterblcoks machined out on a 3 axis CNC mill (HAAS).

Its for a project I'm working on.
 
This first set was a prototype that I wasn't entirely happy with. Surface needs to be lapped and cleaned up. Had to use fly cutter due to some constraints, which is why the contacting surface looks the way it does in the last photo (any why lapping needs to be done).

Specs:
Aluminum 5052
Yes, EK fittings
 

 

 

If you want the quietest fan: Noctua NF-A12x25