Help me optimise WC flow layout...

pre · 2 Dec 2009 at 21:32

Okay, it's been a while since I revised the layout of my watercooling so after some experience from those who are more upto date. :eek:

Components are TC 120.3 installed in the roof of a Blue Typhoon case (a modified Lian-Li 70) shown in the image in green. The block is a D-Tek Fuzion and the pump is a D5 vario with EK Top. The GFX is not being watercooled but is shown in blue whilst the motherboard is shown red.

What I would like to know is what would be the optimal layout - either purple or pink for the pump and the tubing given no reservoir is being used.

Not the excess "tubing" will be a y-splitter used for drainage. For the pink option the splitter would be at the inlet whilst the purple option would be logically placed between the block and pump - although I'm already thinking for the purple option this should be moved to outlet of the block on the way up to the rad....

Any opinions on what would perform best appreciated.

esK` · 2 Dec 2009 at 22:36

Flow designer I used to use that when I first did watercooling.

pre · 3 Dec 2009 at 11:56

esK` said:
Flow designer I used to use that when I first did watercooling.

Thanks but it doesn't really answer the questions I'm asking - and it doesn't even work for a reservoir-less loop!

For example, is it better to have a drain valve before or after the pump? Is it better to have more tubing or less between the radiator and pump? etc etc

Any other thoughts?

JonJ678 · 3 Dec 2009 at 12:34

The two layouts appear to be the same, so I'm a bit confused by the question.

What're you using to fill the loop? It looks like there's a single drain line and no reservoir. This means the drain line has to be before the pump as you'll be filling it through the drain line. It also means draining it isn't going to work as there's nowhere for air to get in.

You want a fill line before the pump and a drain line at the bottom of the case, doesn't really matter if it's upstream or downstream of the pump, lowest point in the case is useful. Then filling is easily achieved, and more importantly you can open both tubes to drain it, air goes in the fill line.

Ocuk sell a valve for about £3 which has both ends threaded, I use this with a blanking cap to finish the drain line and a conventional fill port on the fill line. It doesn't matter if there's a valve on the fill line as water isn't very inclined to come out of it anyway, the blanking cap represents paranoia over the valve leaking (it hasn't yet).

Not convinced I'm understanding you though

wolvers · 3 Dec 2009 at 14:33

I agree with everything Jon says. In addition to that, it won't make a noticeable difference how much tubing you use, just do whatever makes the best, most suitable, layout for your case.

pre · 3 Dec 2009 at 18:52

Hi, thanks for the detailed reply. The fill line was going to be hidden away in the 5.25" bays just after the radiator for both hence I didn't include it.

My concern was whether or not the low flow exiting from the radiator would be a major problem when combined with the drain line causing a drop in flow to the pump and therefore to compensate with more water in the loop after the rad by having more tubing. From what you're both saying it has a minimal effect which was what I wanted confirmation on.

shadowscotland · 4 Dec 2009 at 18:16

without a res - the flow rate is the same throughout the loop (that why people do it

)
One unit of water out of the pump = one unit in (it has no where else to go once all the bubbles are removed)

Monkey Puzzle · 4 Dec 2009 at 18:24

pre said:
Hi, thanks for the detailed reply. The fill line was going to be hidden away in the 5.25" bays just after the radiator for both hence I didn't include it.

My concern was whether or not the low flow exiting from the radiator would be a major problem when combined with the drain line causing a drop in flow to the pump and therefore to compensate with more water in the loop after the rad by having more tubing. From what you're both saying it has a minimal effect which was what I wanted confirmation on.

T-line won't have a big effect on flow at all - better than having a res as theoretically it keeps flow pressure going to the pump.

The only difference seems to be that the purple configuration has the t-line after the pump, which is a bit more fiddly to fill. But the pink looks more cramped - the t-line is a bit bulky and may make getting the t-line fill tube fiddly. When you fill I find it's fairly easy so long as the t-line fill tube is the highest point in the loop, which should be easy enough by laying the case on its side with the pink config, or without moving the case if the t-line fill tube is long enough to bring it to the highest point in the loop. If you have a rad with a bleed screw then you can unscrew it and fill with the fill tube held up high.

Monkey Puzzle · 4 Dec 2009 at 18:26

shadowscotland said:
without a res - the flow rate is the same throughout the loop (that why people do it )
One unit of water out of the pump = one unit in (it has no where else to go once all the bubbles are removed)

It'll always be the same flow throughout the loop; only the pressure changes. The difference between a res and a t-line is you lose more pressure in a reservoir, though it's not a massive amount compared to the pressure drop through a block anyhow.

pre · 4 Dec 2009 at 20:48

Ah okay pressure was the term I obviously wanted.

So this will be the layout I have settled on. It is a bit tight with the t-line at the top but in this position I can have the pump just hanging off the bottom so there should be no vibration of the case.

The drain t-line will be placed after the block instead and angled to the hidden side of the case.

I've been using 120x38mm Panaflos (108CFM/39db/12v model) running at 7v which although quiet are by no means silent). Has the world moved on much in terms of there being good/better 120x25mm high static pressure (rather than raw CFM) fans running in a pull configuration? Or would I be best off sticking with the Panaflos?

JonJ678 · 5 Dec 2009 at 08:59

The idea for draining is to lie it on its side? If so no worries, otherwise it's probably better to have the drain line between pump and block as it'll be at the lowest point in the loop when the case is standing up. I've taken mine from one of the spare holes in my graphics card block. Suspending the pump by the tubing definitely works, I'd feel more comfortable doing so with compression fittings than with barbs though.

38mm Panaflos will do you just fine, don't swap them out for any 25mm fans. 38mm san ace might be worth moving to, but probably not. If you decide noise is important, undervolt the panaflos.

pre · 8 Dec 2009 at 21:54

I seem to have come a cropper with the plan in that the D5 +EK top combo doesn't like the inlet coming straight down from the radiator (albeit through the t-line). Testing it outside a case with the same orientation as in the pic. When the pump inlet is vertical and the outlet is horizontal I seem to get pauses when as great big bubbles of trapped air stop the flow of water. When turning it 90" so effectively the inlet is horizontal and the outlet is vertical, flow immediately speeds up. Anyone experienced this behaviour with a D5 not placed at the bottom of a loop?

EDIT: Ignore I haven't done this in so long the loop wasn't as full as I had expected....