Originally Posted by jwalkermed
Love the red fittings and crystal link. Keep on keepin on.
Originally Posted by nvidiaftw12
Whatever works best for you. I'm not the best at visualizing tubing, so I'll let others suggest if they think another way would be better.
I'll noodle on it.
Originally Posted by Wibble360
I'm with the white compression fittings and dyed crimson tubing.
IMO too many compression fittings / crystal links looks a little untidy - longer arcs of tubing might look more elegant and the crimson would break up the white background of the case nicely
If I used tubing...then there are going to be a LOT of compression fittings in a small area. I'll give it a try and see what it looks like...but some of the turns are sharper than I like with tubing. I don't think it would look long and elegant. But we'll see.
Originally Posted by stren
Red fittings and full crystal link (if you're running just water) with no dye, use tube only on the back side/pedestal. If you use tube, use smaller diameter tube to match the links a bit better. Right now you have some fat tube going on that looks out of proportion. If you run dye keep all the fittings white, however then you may as well have got boring reservoirs so I'd avoid dye.
I had some crystal link in that place for a little bit...but since the graphics card has zero wiggle room...it was tough to get an exact match height wise...and the crystal link wasn't sitting as square as I would like it. If I end up having to keep tubing in that spot, I'll get a smaller size compression fitting. Right now...I've got a lot of $$'s tied up in 1/2 ID x 3/4 OD compression fittings. Easy to change the tubing...not so easy on all the fittings. But...I could spring for a few.
Originally Posted by stren
Also get some GPU backplates powdercoated please
Also frozen q should be able to do you some white end caps for the reservoirs, and the D5's need matching white covers
I have much higher standards for your build than mine
Also I really don't like subloops, I much prefer individual loops with QDCs so you can change parts easily particularly when you have so many radiators and pumps
I've always been thinking backplates....but the look of the PCB has grown on me also. Still thinking about it.
I might inquire about some white tops for the reservoirs.
And believe it or not....I've already been requesting some white D5 dress-kits. Bitspower, and Vincent, have gone silent on me again. Same as when I was first trying to get hold of them months ago. Nothing is more frustrating than simply not getting any response. I'd much rather just get a "no", and I would respect that more, than not to get any type of response at all. Might have to consider those monsoon fittings.....
I can't come up with something that QDC's would help me with. I actually have some. But with blocks on EVERYTHING...I can't come up with anything that I could really do without draining the loop down past the motherboard....even with QDC's. Give me a scenario where they would benefit me. Sell me.
Originally Posted by miahallen
Well, you get some points for creativity, but I don't recommend going with this loop setup. Your pumps are in parallel, which is a no-no (more here
)...you need to come up with a single loop where both pumps are in series.
My suggestion would be: helix res #1 >> pump #1 >> GPUs in parallel (inlet & outlet at bottom for cleanliness) >> bottom radiators >> helix res #2 >> pump #2 >> CPU/mobo/RAM/VRM >> top radiators >> beginning. With everything on a single large loop, the pumps will be working together so you'll have stronger flow, and you'll have a fail safe in case one of them dies.
Oh and yes Chris, you are correct about your assumptions with a closed loop
I'd read Martin's study. It's not directly applicable though. He concluded that parallel didn't have any significant benefits over serial. But I've got both. There are 4 pumps total. I've already got all the benefits of serial pumps. And the pumps for the two subloops are not really parallel, since one pump powers the GPU subloop before being combined with the other subloop. Different than what Martin was testing.
But I will think about the proposed loop.
Originally Posted by m1ndb3nd3r
Ok stupid question time... For Series and parallel loops, is there any testing to show what is better for you's? Is there any tests that shows the difference for radiators I would think parallel would be better as then the water is cooled for a longer time?
I'll probably settle on 2 or 3 possibilities....and then run actual tests for each possibility. Flow and temperature for each. I can't think how running the radiators parallel would cool the longer water. Main benefit of running parallel is to get less pressure drop, ...which should increase flow rate in the part of the loop that is not parallel. It's going to lower the flow rate in the parallel portion...because you are dividing your flow in half. Flow rate won't fall exactly by one half though....because there will be less resistance.