Cerberus-X Watercooling [IN PROGRESS]

Pyrath

Average Stuffer
Original poster
Feb 5, 2016
59
82
Hi all,

Figured here was a good a place as any to document the process of putting an itx build underwater in a cerberus-x. Forgive the first day, as I was too busy being giddy over the new case to remember to take pictures. Will be much more thorough as we go through the process of custom cables, bending tubes, and putting it all together. So for Day 0, we have one 'the PC turns on' picture and the rest will be contemporaneous notes.

For the overall plan, see this fun tinkercad mockup.
Day Zero - Arrival

Unboxing
Pretty standard fare here, with the expanding-foam pads at each corner to provide support to the case, which was wrapped in a plastic covering. If it were plastic or brushed aluminium I'd be worried about scuffing, but the powder coat on these panels feels like it's quite hard. and my fears were unfounded. All the gooey fun bits (handle kit, SFX mounting kit) were put inside the case. I also didn't get the PCI-slot cover that a bunch of other people are worried about, but this has been addressed - see quote.

Yes that was an oversight on our part, those are all mailed out already to everyone that ordered early, you should get them in a few days. My apologies on leaving those out, all cases that shipped this week however have them included...

...For any issues please email me, we'll get any issues addressed and big thank you to all early adopters.

Case
[Todo: glamour shots of all the panels, including the dreaded notch]
Very happy with beige color of panels, not quite the gross smoke-tar-film uv-degraded pukey white from my memories. Which is good. As I said above, the powder coat is robust. It's quite thick and coarse. These things feel industrial or military, and have a satisfying metallic plink when you flick them. Issues - I'm not the only one to note this, but my side panels have a distinct and surprisingly sharp notch (probably from cutting) along their edge. To be honest I only looked because I knew to look for it from a previous thread.

After removing the panels, the interior frame was laid bare [Todo: frame pictures]. Very nice black everywhere, smooth matte finish. Rounded corners! No sharp edges (and I tried to find some). Overall really high quality. However, there are always complaints!

Firstly, I was missing a screw(!) in my case. It comes assembled with many countersunk holes already filled with these delightful tiny little screws. So I used one of the screws provided to make sure my case wasn't going to fall apart on me.

Second, I am worried about the amount of flex in the rear of the case before IO panel, rear adapter (psu or no psu) and GPU are installed. Specifically at the top end of the PCI slots, right underneath the IO/rear adapter. My case was a little bent there out of the box. Perhaps a target for revision or packing support. It's fine now that there is hardware installed which add rigidity to the rear of the case.

Third, some of the the ball-pin things are loose, but they can be hand-tightened back into place. I don't expect them to cause any issues but it's a bit of a downer that they are so loose.

That's all I can remember now... taking requests.

Building [todo - LOTS OF PICTURES]
Due to the lack of sharp edges, ability to remove all the panels with ease, and the lack of any top panel in the frame, building in the case was really quite easy. But I'm new to the SFF world, and some lessons were learned.

Lesson 1: If using the SFX PSU adapter which includes the short fancy power cord, you must first thread the entire cord through the rear cutout. I had it all setup before I realized I had done that out of order.

Lesson 2: Install the handle plate / handle last. If you do what I did in lesson one, you'll need to remove the cord from your PSU and pull it out. TURNS OUT, you cannot do this when you have the handle plate installed. The gap between the SFX PSU and the handle plate is too small to remove the cord.

Lesson 3: Those tiny ass screws are too damn small. I consider myself pretty diligent, but I lost one, and because I had to use another one to fill in the missing frame screw, I was then down two screws. They don't give you extra, and as of right now I can't screw in and secure the rear power plug.

Lesson 4: I knew this and ignored it when planning, but the motherboard standoffs are very firmly affixed to the tray. I'm using an ITX build and don't really need the extra ones. if they start to get in the way of radiators and what-not, I'll have to try and rip them off. I hope it doesn't come to that. Remains to be seen.


Day 0 End
I'm very tired and wanted to play some videogames with friends tonight. I expect to have time for fitment and a dryrun of watercooling this weekend/next week. Will take loads of pics then. For now I leave you with this (shoutout to @MarcParis who loves LEDs). The CM Hyper 212 is hilarously large, and until I set up all the waterblocks I won't be able to close the side panel. Forgive the dust, this guy has been sitting on a cardboard mat on the floor of my room since June waiting for this case. I'm so excited!

Day 1
Hoo boy, am I glad you guys provided a CAD file for me to test fit things... It's quite snug! Today was fun, as I spent the afternoon shoving all the watercooling bits inside the case. Additionally, removing stock heatsinks, throwing liquid metal on everything, and slapping on the waterblocks. Haven't tried turning it on yet because it's near midnight and if I see a problem now I'll be up until three fixing it.

Actually I'm super exhausted so instead of a novel like Day 0, have an imgur album.

Once again building in the case is a pleasant experience since it is so dang open. I've added several shots of how bendy the panels can be - check the WIP imgur album. I can only hope that work fatigue of the powder coat / panels themselves is not something I will have to worry about during the lifetime of the case.

Todo:
Custom length/sleeved cables, either myself or paying to avoid the hand cramps.
Find a few missing screws for GPU backplate and case power
Wash out rads and blocks
Bend tubes (hurk)
Leak test
Acquire side panel window (lookin at you @KSliger)
Finished?



Day 25??? Ho Ho Ho!
So the end of the year always ends up being a flurry of work and I have little to no time for recreational pursuits like this monster. A few spare moments in the last few weeks were spent bending, but I ended up throwing most of my gear in my car and taking it to the parents' house.

The Todo above has been remedied - mostly because I was really tired of having no primary pc to putz around with. So all the dreams of custom sleeving and nice bends and what not are delayed! I plan on redoing everything with nicely sleeved cables and cleaner bends after we see how great the noctua sterox fans are in january.

With these ugly bends and only 2 fans working out of three (I forgot to bring my fan cables!), this thing is dead quiet and I have yet to get my GPU over 40c (ambient temp ~27). Merry Christmas!


 
Last edited:

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
Thanks for sharing your experience..:)

You are right...i love leds...:)
It’s well known that i consider rgb and tg as evil...lol

Ps : about flex in the case, well, that’s why personally, i always put the side bracket on. When side panels are on, flex is largely reduced. It’s common with a lot of cases..:)
 

Arboreal

King of Cable Management
Silver Supporter
Oct 11, 2015
818
814
Good to see a Cerberus X out there and being built.
Now, there seems to be a small conflict here...
Should I nominate it for the Space inefficiency thread - ITX in an ATX CASE, or does the TOO BIG TO FIT EVO 212 cancel that out? :p
Looking forward to see how the waterworks are implemented
 
  • Like
Reactions: Pyrath and cmyk78

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
Space inefficiency is you go itx with aircooling..;)
If it’s fully watercooled, it will be become oversized cooling..:)
 

Arboreal

King of Cable Management
Silver Supporter
Oct 11, 2015
818
814
Space inefficiency is you go itx with aircooling..;)
If it’s fully watercooled, it will be become oversized cooling..:)

Well put!
As the Cerberus X is smaller (in volume) than Fractal's Define Nano S @28L I can't really call it inefficient. I have no experience in watercooling, and not much interest in trying it.
That said, more radical cooling of high end kit in a great case is always interesting and entertaining to see.to
 

Pyrath

Average Stuffer
Original poster
Feb 5, 2016
59
82
Finally had some time to putz around. Updated the OP with the below post!

EDIT: Apologies for the shit pictures, once the sun goes down I have no decent source of diffuse light

Day 1

Hoo boy, am I glad you guys provided a CAD file for me to test fit things... It's quite snug! Today was fun, as I spent the afternoon shoving all the watercooling bits inside the case. Additionally, removing stock heatsinks, throwing liquid metal on everything, and slapping on the waterblocks. Haven't tried turning it on yet because it's near midnight and if I see a problem now I'll be up until three fixing it.

Actually I'm super exhausted so instead of a novel like Day 0, have an imgur album.

Once again building in the case is a pleasant experience since it is so dang open. I've added several shots of how bendy the panels can be - check the WIP imgur album. I can only hope that work fatigue of the powder coat / panels themselves is not something I will have to worry about during the lifetime of the case.

Todo:
Custom length/sleeved cables, either myself or paying to avoid the hand cramps.
Find a few missing screws for GPU backplate and case power
Wash out rads and blocks
Bend tubes (hurk)
Leak test
Acquire side panel window (lookin at you @KSliger)
Finished?
 

Nanook

King of Cable Management
May 23, 2016
805
793
I’m planning a similar build as you. With all the water cooling components on the bottom half of the case. :)
The panels on my C-X are bendy as well... there are slight gaps when the panels are assembled.
 

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784

I just realize that you are considering corsair ML 120 fan on HW labs radiator, why?
Those fans are not really air pressure oriented whereas hw labs needs pressure due to their high fpi..:D
 
  • Like
Reactions: Pyrath

Pyrath

Average Stuffer
Original poster
Feb 5, 2016
59
82
HWLabs GTS aren't terribly restrictive. MLs are a good compromise of sound and airflow. http://thermalbench.com/2016/07/12/corsair-ml120-pro-120-mm-fan/3/ I personally think static pressure is kind of a bunk and inconsistent statistic. We'll see how they do. I was hoping Noctua's new Sterox 120mm would be out by now but they've been pushed to Jan. If they're massively more performant, I might have to buy some more fans next year.
 
Last edited:

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
Well, based on thermalbench review, i've read more than once...and graphs could be misleading.

I found this graph really useful, because i rely more on RPM than their DB measurement. (ML fan start to be noisy at around 1400 rpm, Gentle typhoon are still pleasant at 1600 rpm but become loud by 1750+ rpm)

But look back at the graph, on 1400rpm and lower...ML fan are way behind Vardar (equivalent to Gentle typhoon or Fractal Design HP-12)...thus ML fan could perform great but at 2000+ rpm (so crasy loud...:)).

You are right, I'm putting a lot of hope in upcoming Noctua NF-A12 (sterox blades)..but i fear that this fan will only spin up to 1500rpm...and I don't expect to perform as good as Gentle typhoon/Vardar/FD HP-12, at least, on radiator.
Moreover, Noctua motor tends to be noisier than Vardar/Gentle typhoon at high RPM (ie 1500+ rpm)...my experience on Noctua PPC at 2000rpm was painful & noisy...I prefer by far my Gentle Typhoon noise...even at higher rpm..:D
 

phazeshifta

Cable Smoosher
Nov 23, 2017
11
1
It's funny you keep saying that the ML 120's aren't great for Hardware Labs, but that graph is directly contradicting the information you keep posting. They perform better than everything but the Sliverstones and Vardar's at almost every rpm level. Do you have something against Corsair that makes you keep posting information contradictory to the very graphs you are taking as gospel?
 

Pyrath

Average Stuffer
Original poster
Feb 5, 2016
59
82
It's funny you keep saying...

I think what @MarcParis means when he says ML's aren't 'good' is that they aren't the best possible fan given this use case. And according to that graph, at low RPM's (which is where I'll be using them) he is right. In this community oftentimes 'not the best' can be 'bad'. No harm done by Marc.

Vardars are the better choice by that graph. I've read and heard some interesting anecdotes about the ML noise profile being different/less obnoxious than other types of bearings, and I wanted to see for myself - so that's my reasoning for grabbing not 'the best' low RPM fan. I expect to be able to run them faster than others. We'll see.
 

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
It's funny you keep saying that the ML 120's aren't great for Hardware Labs, but that graph is directly contradicting the information you keep posting. They perform better than everything but the Sliverstones and Vardar's at almost every rpm level. Do you have something against Corsair that makes you keep posting information contradictory to the very graphs you are taking as gospel?
Euh...are you reading properlly the graph or have you read my post?
At 1000rpm, ml120=100fpm, vardar=200fpm...
At 1250rpm, ml120=150fpm, vardar=225fpm...

At 2000rpm, ml120=vardar=325fpm...

...
 

Pyrath

Average Stuffer
Original poster
Feb 5, 2016
59
82
Minor Update - turned it all on to make sure it'd boot - and it does! Woo!

The ML's are remarkably quiet, but can move quite a bit of air (and make quite a bit of noise) if you let them run wild at max RPM.

stole borrowed a heat gun from a coworker today so will perhaps attempt some bending this evening.
 
  • Like
Reactions: MarcParis