JayGee's Dark Base Pro 900 Water Cooled Build

Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
would use the green route, phanteks state to go through the inlet port , but technically will be fine going via the output port for the bottom card . Really couldnt when you use alphacools extremely restrictive GPX blocks which are do give a negative flow effect when reversed
Aren't the green and red routes logically the same - both cards get fed through their designated inlet ports and exits through their designated outlet port?
Oc6LE1l.jpg k0nRWV7.jpg
Just trying to figure out if I'm missing something (this is only my second water cooling build and the first where I intend to feed two water blocks in parallel).
Blue arrows to show the flow through/between the parallel sections.
 
Soldato
Joined
1 Dec 2015
Posts
18,514
option 2 :) looks neater - im sure from using phanteks blocks it was the left port that was ' in ' and right was 'out'
LKivDk4.jpg

but again shouldnt make a difference in performance if its hooked up like your diagram

though i could be wrong, Ti may have a different flow path- the arrows on the block to state flow path is shocking!

I believe the left side, flow travels to the jet plate , then through the fins and out or the right side

m9VT4HF.jpg

as you can see block traps bubbles... best cure, different pump speeds and looks of flipping the case !
 
Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
option 2 :) looks neater
That's cool, then I think that this is mostly a question of aesthetics, and being realistic that means that the final routing won't really be figured out until the heat gun comes out and I can see how things look in the flesh. :cool:

im sure from using phanteks blocks it was the left port that was ' in ' and right was 'out'
I'll check when I get home this evening but I'm pretty sure that the inlet ports are closest to the pump end in my build.
EDIT: Picture of the ports showing indicated direction
ZneCSpQ.jpg

So are you saying you double tubing the GPUs?
Sorry, I'm not familiar with the phrase "double tubing", but if it means running the water through the two GPU blocks in parallel, then yes.
If it helps here's a (badly drawn) schematic of the proposed loop to illustrate:
9xeLlAL.png
 
Last edited:
Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
hmm... I only used one tube, would I really notice much difference running two?

Thanks,

Stelly
That's probably a question for the water-cooling section of the forum rather than for my project log, but for what it is worth, and please bear in mind that I'm no expert (my "wisdom" is only what I've gained from watching stuff on YouTube and reading various forums)... it probably won't make a great deal of difference in most cases.
 
Associate
Joined
2 Jan 2018
Posts
191
running the GPU's in parallel looks better to my eyes personally

it's a regularly debated question really... series vs parallel... and really depends on your system and how your specific cards behave... the only way to know is really test both and see what the results tell you for your use case scenario.
 
Associate
Joined
10 Dec 2017
Posts
49
Water like any liquid follows the path of least resistance so which way it goes can't actually make that much of a difference. Am I thinking wrong here? From a physics standpoint. I don't think relative motion of h2o resistance plays a factor in the actual waterblock?
 
Associate
Joined
2 Jan 2018
Posts
191
nope, you're right... it's down to pressure at the end of the day... parallel lines reduce coolant pressure, where a series loop keeps constant pressure over the cards

if you've got a pump like a DDC that can handle the extra pressure then parallel or series loops are open to you... but if you're just on a low pressure / constant flow rate pump like the D5, then series is a better option
 
Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
Busy day, I'll add some pictures either later this evening or during Sunday (now done), but for now a few words on what has been done today:

1. Stripped everything out of the case frame.
2. HDD frame primed and repainted (one of the first things I did to the case was to dremmel out the HDD supports, I'd then fitted it back to check fit with other components in it's rough state).
3. Flushed out both rads.
4. Measured up final placement of the rads and then drilled out and filed additional mounting points for the top rad (pre-cut mounts only fitted with 4 out of the 12 mounting holes on the rad).
5. Fitted front rad with all three fans.
6. Fitted top rad using the new mounting points.
7. Figured placement of and fitted in-line temp sensor and flow meter to the far-side outlet of the front rad.
8. Got the first piece of hardline in - the long, near-side, drop from the top rad to the front rad that needs to skirt around the res and pump.
9. Mounted up the res to the front rad and also the pumps to the front rad (fitting to the bottom mounting points interferred with the ports on the rad).
10. Fitted to the EK multiport top to the EK X3 res, and of course the top ports and bottom ports didn't align as I wanted them so I had to sand down the top of the tube until I could get the top to screw down to where I wanted it.

Plan for tomorrow:
1. Refit the HDD frame so that I can measure up, make a backplate.
2. Figure out SSD placement and fitting on the backplate.

Stretch goals for tomorrow:
1. Hardline work: [a] res-to-pump; flow-sensor-to-res; [c] vrm-block-to-cpu-block; [d] inter-gpu-block-connections.
2. Refit the Mobo and graphics cards.

EDIT: Pictures Added

Repainted ex-HDD Frame:
9bqonff.jpg

Stripped out components:
7opN7ci.jpg

Couple of shots of the res, pump, rads and the first piece of hardline (taken in the bathroom to try and get some decent light!):
p40hCtj.jpg
MhcO6Vx.jpg
 
Last edited:
Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
Plan for tomorrow:
1. Refit the HDD frame so that I can measure up, make a backplate.
2. Figure out SSD placement and fitting on the backplate.

Needs some finessing but otherwise I'm happy enough with this:
muJjOXa.jpg

Stretch goals for tomorrow:
1. Hardline work: [a] res-to-pump; flow-sensor-to-res; [c] vrm-block-to-cpu-block; [d] inter-gpu-block-connections.
2. Refit the Mobo and graphics cards.
Didn't get anywhere near the stretch goals.
 
Soldato
Joined
1 Dec 2015
Posts
18,514
That's cool, then I think that this is mostly a question of aesthetics, and being realistic that means that the final routing won't really be figured out until the heat gun comes out and I can see how things look in the flesh. :cool:


I'll check when I get home this evening but I'm pretty sure that the inlet ports are closest to the pump end in my build.
EDIT: Picture of the ports showing indicated direction
ZneCSpQ.jpg


QUOTE]

You can see in the pic and Phanteks aweful arrow system, left side is the inlet - the one that has the arrow facing up, and purport is the right... The one with the bloody arrow facing down and into the block hahahah , honestly that's a really badly done thing right there !
If running series then your outlet from bottom card will go to outline of top card and out the inlet.
If you ran parallel , you can go from the bottom left port of the card. Two tubes at the top of the bottom card going into the bottom of the top card, then single exit tube at the top right side port into the VRM. You have dual D5 pump so pressure should be fine :)
 
Soldato
Joined
1 Dec 2015
Posts
18,514
So the arrow pointing towards the block isn’t the inlet, and the one pointing away from the card isn’t the outlet. Damn, that is screwed up!

Yep, just double check the install manual as it will state there , but confused me at first the arrows . You can see if travels in on the left, hits the slit, then is forced left and right at the fine grid underneath and travels around the card /vrm to the outlet.
Again in theory shouldn't make much of a difference, the block and jet plate/fins are restrictive like Alphacool GPX block or CPU blocks so should be fine
 
Associate
OP
Joined
3 Jan 2018
Posts
205
Location
Cambridge
madness isn't it. you'd think the Out symbol is the damn inlet! haha

so if you went series you'd go out the exit of one into the exit of another, if you run parallel with inlet on the bottom card from pump and top exit on the top card to VRM- should be good
Yep, I’m cool with that, my current rig has water cooled duel GPU’s (they’re linked serially using an EK block).

For this new build I’m set on running the two GPU’s in parallel (mostly because of the aesthetics). The debate I’m having with myself is about whether I can find a creative way to run the hardline that will be both practical and pleasing on the eye. That said I’ll probably simply do as you suggest: in bottom left and out top right - KISS. :)
 
Back
Top Bottom