help me decide on loop setup and other WC tips

[MadMan007]

Thanks for the input, that's good to hear about the pump. Maybe in the open it's loud but it will be inside the sound-deadening P180 anyway, and mounted on some of Petra's Gel stuff with bolts isolated by grommets. Shouldn't be too loud then I hope

I was not really considering push-pull, just one or the other so seeing as there's no difference it will let me test both airflow layouts. A difference of 1.5C doesn't make going push-pull seem worth it to me.

I think I'll be good for mobo cooling too. I did make a thread in the Intel mobo section though outlining my reservations after remounting the heatsink assembly though; shouldn't have done maybe I'll get some thermal adhesive heh, or get some nylon bolts for hardmounting instead of relying on the weak springs...ah well.

 

[BobcatDan]

I have an MCP655-B, and I have no idea why people say it's loud. I do have it mounted on Petra's gelstuff pad, but with the case open for bleeding I had to have me ear a few inches from it before I could hear a subtle high pitched whine. If you had it hard mounted, I could see it making a lot of noise, because that kind of high pitched vibration would turn the bottom of your case into a speaker, but with a decent isolating pad you should be fine.
Dan

 

[Otter]

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 ...

Your temps were slightly lower with the 355 than the 655?

 

[Otter]

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?

Yes. The centrifugal pumps we use will not pump gases at all. Hence, if you get air in your impeller chamber, the pump can't pump it out. Add that some pumps will shred themselves very rapidly if they run dry, and the pump is the absolute worst place to have air in your system. With the reservoir or T-line right before the pump, you can be sure no large bubbles will reach the impeller chamber.

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?

This may vary from setup to setup and may be different even in the same setup with the fans at 5v and 12v. But in general, push will probably give the best absolute thermal performance, and pull will give the best noise/cooling ratio. The differences won't be large either way, though.

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?

If you don't have a quiet PSU and quiet fans and you haven't done anything to silence your hard drives, you probably won't notice the 655 much. OTOH, once you've optimized the rest of your system to be as silent as possible, the 655 will be very noticeable. It would blend right in with a high but constant noise system, though.

In contrast, the 350 is barely audible even if you're standing right next to it with no other noise sources in the room. It makes just a quiet hum. The 355 is slightly louder than the 350 and higher in pitch.

Is the MCP350 or 355 really practically the same performance plus flexibility to go quiet?

Assuming you got rid of the elbow at the inlet of the 350 with the DIY mod or Petra's top, yes. Most of us here tend to obsess over minute differences in performance, but once you've got about 1gpm flowing through your loop, more flow won't reduce your temps by much. Look at this chart for D-tek's Fuzion CPU block.
http://www.dtekcustoms.com/ProductImages/waterblocks/fuzion/TChart.gif

Going from from 1gpm to 2 gpm, the pressure drop more than triples, but you get only about 15% better thermal performance. And in a typical loop, you won't get anywhere near twice the flow from the 655.

 

[QuietIce]

Your temps were slightly lower with the 355 than the 655?

Slightly, ~1°C less with the MCP355 stock. With the Petra's top it got worse, only ~0.5°C less than the MCP655 @ P5. Go figure - the numbers surprised me too.

And in a typical loop, you won't get anywhere near twice the flow from the 655.

Quite true! The MCP355 and MCP655 @ P5 had near equal flow - at least close enough that they were within the error of measurement. But I'm not sure you'd call my test loop "typical" ...

 

[MadMan007]

Thanks for the detailed replies Otter, and input by others I already have the parts and wouldn't exchange them unless there was a serious issue. I am not going for 'absolute silent' just 'nice and subdued quiet,' a little fan air noise or minor pump hum won't matter. I should have asked about the 355 first though, it is much more compact and easier to place! No matter now...

So, I am probably going to go with a t-line, my picture #1 looks like a good loop sequence ?

 

[QuietIce]

 

[MadMan007]

OK bump for a few more questions:

1) How do I 'prime' the pump? D5 directions are very strongly worded regarding running the pump dry. Is it just a simple matter of providing some fluid at the intake and turning it on or is there some better special method?

2) How do you guys go about measuring and laying out the tubing? I'd like to leak test separately outside the case before putting it on components. Of course in order to do this I have to pretty much mount everything in position, measure and cut the tubing, clamp it etc. I guess this is the only way to get the loop made with correct tubing lengths?

3) General rad/waterblock/system prep -any special cleaning needed or recommended? I'm using distilled water+water wetter+iodine and am not sure that prepping with some other potentially contaminating material is a good idea but if there are releasing agents or other stuff on the components, tubing or rad I'd like to clean them up properly. I've seen vinegar mentioned but then also not recommended so I'm just not sure here.

btw I'm just using basic Masterclear tubing.

 

[QuietIce]

1. Priming the pump basically means being sure there's water at both the inlet and outlet. The water from your T/res should enter the pump and leave the pump by gravity before you turn the pump on. Even when you see water up the tube from the outlet listen to your pump - it'll whine a little if it starts getting dry or big bubbles run through it. If you've got the variable speed 655 use P2-3 for awhile before kicking it up.

3. I've always used a little white vinegar in the rad before using it. Just let it sit in there for 15-20 min (I've even gone overnight) and rinse w/lots of tap water followed by one rinse of distilled.

2. Man, you've opened a can of worms with that question! I don't see any good reason to put a loop together outside the case for leak testing. When you put all that stuff in the case you might loosen something and have a leak, so why bother???

If you just want to put it all together to play with it and see how it's working that's cool - I did the same thing the first time! Just rough cut your tubing (a little long - you can trim later) and put your loop together. I like to use a large clean bowl/pan for a "res" (since I use T-lines and they're a pain outside the case). Put a tube into the bowl then pump>blocks>rad (order doesn't matter here) and put the final tube back into the bowl (I use all the left-over from cutting for this to make it extra-long). Use distilled water and fill/bleed.

I like to hold the final tube down low in the bowl and leave the "inlet" tube just under water level (don't want air back in the loop). Debris will come out of the loop as the water is running, you want those to settle to the bottom and stay in the bowl. That's why I like the loop outlet low and the inlet high - preferably in opposite corners ...

 

[MadMan007]

Thanks for the priming tip, I'll make sure there's liquid through the pump before turning it on.

Ok I'll clean the rad. What bout the blocks and tubing? Keep in mind all this stuff is new.

Any help on point #2? I don't really fancy assembling and positioning everything to get the tubing right, taking it apart to leaktest then remounting it. But if that's the way to go so be it.

 

[QuietIce]

You don't have to mount everything to get the tubing more or less correct. Lay the case on it's back and put the MB in - maybe two screws so it won't slide around - then add the vid card, no screws. If you want you can set the CPU block in the CPU position (w/out the chip). From there you should be able to get w/in ~3" long of the correct length and you can trim it later when you've got everything mounted in the case.

Maybe I'm missing what it is you want to do. Are you trying to assemble the final loop outside the case and then mount it w/out taking the tubing off the blocks?


BTW - If you're going to assemble the loop outside the case you can rinse it out by simply running the pump and keeping the outlet tube away from the inlet tube as above. I still recommend cleaning the rad separately but put it in your external loop afterwords ...

 

[MadMan007]

I'm stupid and hadn't thought of simply mounting the hardware and running the pump on its own That will work best since the tubing lengths will be right and I won't have to juggle the parts to mount it.

I've prepped all the pieces and will finish modding the case and mounting it all tomorrow, if not Monday.

 

[Torlough]

First this is a great thread for a first time installer, it has answered alot of my questions, qudos to MadMan and all those who answered.

I am doing a first time set up. I have been told to do both a T line and res but it sounds like I only need a res correct?

Also i was told to clean the components with an alcohol/ water solution before use, then rinse with distilled water. What percentage of Alcohol should I use with the water??