help me decide on loop setup and other WC tips

This is kind of a continuation but different topic to my other thread. So within the next few weeks I'm going to set up my watercooling. MCR320 rad, Apogee GTX CPU, MCW60 GPU, MCP655 pump. Yate DS12 47CFM fans on the rad (F3-F5) and Scythe 67CFMs (F1 and F2) case intake. Maybe a micro-res. My quandry is that I'll be putting this in a P180 which is a little short on the frontal intake for the main mobo bay area. I've come up with some drawings of layouts:





There aren't many option for mounting the pump unfortunately, any where lower below the mobo would interfere with cards or cables too much. I may hack the case up a bit by cutting down the area that says 'antec design' to move the pump lower and more out of the way of the front fan. Even as it is though the front fan isn't too blocked.

First off for loop layout some input is appreciated. I was trying to do pump->rad->CPU->GPU->(res/tline)->pump is that the best sequence? What's the best loop for the t-line and/or res? I am leaning toward the second picture atm with the res in the back of the case on a swivel to get at the slots, it has the width to fit there. Will it be ok to have the res low in the loop or will that cause problems filling?

Second is the overall airflow. I know that ideally the rad would have fresh air intake however doing this means F1, F2 are outputs and that means little airflow toward the non-watercooled vid card components. So what I was thinking is the rad fans are output and F1, F2 intakes. I would make up some kind of duct for F2 to make sure it's not recirculating hot rad air. Either way it's not great because one of the regular case fans is not going in the best direction.

Fan types - 3 Yates versus 2 Scythes is pretty close to balanced in terms of air flow. Plus I already have them but I'm not against buying others. Would Scythes on the rad make a noticable difference? I would probably undervolt them anyway to balance the airflow and possibly for noise reasons even though the Scythes are pretty quiet so I'm thinking the Yates will be fine for the rad.

Also what's the best way to prime the pump? The pump instructions are very emphatic about not running it dry and priming it. Is just having the intake hose filled with water enough?

Can't you shove the rad fans on top of the rad?

Maybe but this layout is better to keep it as externally compact as possible and also use the rad as a bit of a sound muffler. I don't mind losing the top 5.25" drive bay at all and I don't think it will make a difference in performance? Does push or pull through the rad matter?

I think pulling is better. Don't quote me on this though

place the fan on top of the case but underneath the rad. It doesn't matter if you go with a res or a t line. i prefer res over t for some reason. However, you should pump water directly from pump to the water block instead of to rad. It will give better flow to the blocks heat will get moved out quickly.

pump => CPU => GPU => Rad => to Res or Back to Pump (if u got a t)

EDIT:

You can have the inlet hose filled up with the water but you will need pour more water in till the loop gets filled up so there is steady flow of water to the inlet of the pump. As for your case airflow problem, what you can do is cut a hole on the bottom of the case and install a fan with a filter and let cool air from bottom of the case come to top.

Another problem you need to tackle is the PSU. Once you install that PSU you will not have room for the tubing. What you can do is move than rad past the end of the case (the barb part of the rad will be outside of the case over the psu) from there you can route the tubing from outside to inside

I had thought it was better to put the rad after the pump to minimize the pump's heat dump in to the water but I guess not. If going to the CPU first would be better that's easy to do, it will mean more tubing though because I will have to go from the GPU up top to the rad.

Maybe you aren't familiar with the layout of the Antec P180. The shot I took is only for the mobo area, there is a separate area below which you can just see the top of for HDs and PSU.

i see...i didn't see that

pump->GPU->CPU->rad->res/tline->pump would be a shorter layout and put the waterblocks right after the pump. I wanted to do rad->CPU because I figured that way the CPU gets fed water fresh off the rad but I may be thinking of how a loop works wrong...if the water equilibrates after some time anyway then it wouldn't matter yea?

it doesn't really matter which you put first (GPU or CPU). There will few degrees differential between the CPU and GPU.

Even though I have no problem with the rear fan reversed (I run two of mine that way ) you might think about flipping the direction of all the fans. Have the rad fans pulling in and the other two pushing out. That way your case gets relatively cool air (the rad doesn't heat air much at all) and the rad gets ambient air instead of air that's cooling the chipsets. The air vent below the rear fan will equalize pressure and ensure there's air flow around your vid card, so I wouldn't worry about an extra hole, either.


The loop flow of res > pump > GPU > CPU > rad > res makes a lot of sense for the component arrangement you have pictured in #2 ...

I would keep the res as high up as possible. I hear that helps with bubble issues.

QuietIce said:

The loop flow of res > pump > GPU > CPU > rad > res makes a lot of sense for the component arrangement you have pictured in #2 ...

Click to expand...

Ok so another vote for having pump->waterblock instead of pump->rad->waterblock ?

Oh yea, another thing. This is for a quad core and a single graphics card (8800GTS atm, small possiblity of a GTX in the near-ish future if not a new single card when they are released) this setup should be enough for that? Would stronger rad fans help more for the heat load?

I will try reversing the fans both ways I guess. With the front door open on the P180 the front fan can either exhaust or intake pretty well, with the front door closed though it may be a different story. Guess I'll just have to test and see.

Since the coolest your liquid is going to be is when it exits the radiator should not that liquid then go to the cpu to cool it.

Since liquid is not compressable and the system is closed and pressurized, there can be no difference in flow once the system is closed.

Therefore, it would be logical to direct the coolest liquid to the priority item to be cooled in the system.

Alrighty well I've somewhat decided, ahead of time without testing though, to go with a t-line (no real difference versus a res? Filling with a funnel isn't a problem at all ) because the layout would be more easier for getting to the mobo rest of the computer. So the t-line should be at the pump intake for priming purposes yea?

I also think I'll try F1, F2 as intakes and the rad fans as outputs. Getting fresh air through the rad would be nice and I'll still try to reverse it but the front fan pushing air out will be weak I feel, the drive cage will block it from drawing air plus the P180s door isn't really made to have a fan pushing air out imo. Or maybe it would be ok we'll see

Another thing, I got some 1/2" ID 3/4" OD inexpensive (not Tygon) tubing along with the other final bits delivered yesterday. Tested out with a bare barb and it seems to slip on somewhat easily. I know with clamps that leaking shouldn't be an issue but is 7/16" ID over 1/2" barbs much better or recommended? The sticky testing this doesn't show much performance difference just the security of the connection is what has me concerned a little.

There is very little difference in performance between 1/2" and 7/16" ID tubing. The smaller tubing fits much more tightly, but neither will leak if you use appropriately sized worm drive hose clamps.

As long as you put your pump right after the res or T, component order doesn't matter too much. Multiply the watts added to the coolant by any component by .0038 and divide by gallons per minute and you'll get the temperature differential between inlet and outlet in Celsius.
c=.0038*watts/gpm

One question: Quiet seems to be fairly high on your list of priorities. Why are you using the MCP655 instead of a modified MCP350? If you already have the 655 it makes sense. But if not, you might want to give the pump some more thought.

Otter said:

One question: Quiet seems to be fairly high on your list of priorities. Why are you using the MCP655 instead of a modified MCP350? If you already have the 655 it makes sense. But if not, you might want to give the pump some more thought.

Click to expand...

If it's the "Vario" version of the MCP655 (adjustable like all the old ones) then a P3 setting is probably all that's needed and it's extremely quiet. If extra cooling is needed for a few hours P5 is always there to drop a couple of degrees ...

Well, yes, but the DDC3.1/MCP350 will be that quiet running flat out and will push enough coolant to get you very close to the same load temps as the D5/655 would give you on the highest setting.

Otter said:

Well, yes, but the DDC3.1/MCP350 will be that quiet running flat out and will push enough coolant to get you very close to the same load temps as the D5/655 would give you on the highest setting.

Click to expand...

That might be true, at least, I have no direct eveidence to say it isn't. Due to a mix-up I ended up with an MCP355 instead of an MCP350 to test (I already had a 355 ). But the stock MCP355 only averaged ~1°C lower (CoreTemp @ SETI load) than the MCP655 @ P5 ...

Otter said:

There is very little difference in performance between 1/2" and 7/16" ID tubing. The smaller tubing fits much more tightly, but neither will leak if you use appropriately sized worm drive hose clamps.

As long as you put your pump right after the res or T, component order doesn't matter too much. Multiply the watts added to the coolant by any component by .0038 and divide by gallons per minute and you'll get the temperature differential between inlet and outlet in Celsius.
c=.0038*watts/gpm

One question: Quiet seems to be fairly high on your list of priorities. Why are you using the MCP655 instead of a modified MCP350? If you already have the 655 it makes sense. But if not, you might want to give the pump some more thought.

Click to expand...

Ok well I've already got the 1/2" ID tubing and I'll use clamps of course so I'll just keep that.

Did you really mean 'pump right after the res or T' or 'right after the rad' ? Do the res or t-line even affect temps?

Oh yea, any comment on 'push' versus 'pull' for rad fans? I'd always thought push was better to have more air pressure on the fin assembly, that's how I always did aircooling, but is WC different or is one known to be generally better?

Quiet is important but not HUGELY so, I'm not going for 'silent' just muted and sound quality matters too. A high but smooth noise floor is ok. Can anyone give me any kind of idea of the relative noise levels? Honestly I got the MCP655 just because it was usually mentioned as 'the best' and it is the vario speed model but I didn't have one before, the one I got is still 'NIB' and could be exchanged or something I suppose. Is the MCP350 or 355 really practically the same performance plus flexibility to go quiet? I see it has higher pressure but lower flow rate (please no one go crazy on my likely misuse of terminology ) It would be nice to mod the MCP655 so I could tune the speed from outside the case...hmm

I'd listen to Otter.

Order don't matter much, as long as after the res isn't a block, I remember it being something about reduced flow. In fact, I used to use a res and now I'm on a T-line and I love it. Even more compact, easier to deal with, just takes a little longer to bleed but I have no problem with that... it's only a one time issue per 6 months or so. The res just slows down flow a little, and fast flow = better performance.

I have a push-pull configuration on a MCR220, connected to a baybus fan controller. The difference with JUST push, vs JUST pull is none. The difference between JUST PUSH vs PUSH-PULL, about 1.5C. That's after alof ot reconfig-ing, ambient temp watching, yadayada keeping variables at a constant to really figure out the difference... and my conclusion is that the difference is only abouy 1.5C.

What really matters in quad clocking, at least from my experience, is ambient temps, northbridge watching... with temps and voltage in mind, and some active cooling (i just used the fan from EVGA's 680i package, which is more than I can say for ASUS's P5N32-E... which even with the optional fan was 10C higher!) With the northbridge temps under control, I can easily push my q6600 in the 3.8ghz range, and that's before the vdroop mod I'll be doing tonite.

What clamps are you using? Some advise for some advise against, but I am for the metal clamps... i feel alot more at ease with it... course I'm also on a 7/16" ...

STICK with the 655. I could barely hear it as it is. You'll hear your fans before hearing that pump, great performance, super quiet.