Help I'm under attack by the tinker bug!!!
 

Got some questions but first some setup to explain....

Looking to rearrange/clean up some of the stuff for my CPU loop in my 'concoction' My setup for now
So that I can get a little cooler and deal w/ some (unplanned :mad: ) airtraps. Will swap my 1250 for an MCP600 for starters.

Currently flow path is from the pump/res -> CPU ->rad(top) to res/pump(bottom) to CPU. Will reverse this so as not to fight gravity flow-wise.
Thoughts please???

Want to get rid of the res for CPU loop also. Just saw a post here last night where there was a bleed tube sticking out the top tank of a heatercore (in essence it becomes the res/airtrap (on purpose:D ). Anyone else do this?
More thinking please???

I'm actually having trouble trying to work out what goes where on that picture (you have me scratching my head over here).

But to avoid air traps I would put a res at the top feeding the pump inlet via gravity - this should also act as a nice airtrap if you plan it right as a well as an easy way fill the system. that place above the rad looks like a nice place to put in a res with a fill point sticking out of the case at the top.

Maybe if we saw the case without the wiring in it - just the pumbing it might be easy to see what goes where, I'm sure a couple of those tubes end in dead ends :D.

~ Boli

This is the basic idea
 

How about this?

http://home.comcast.net/~mikeehrlich...ulooptest1.jpg

Water is feeding into pump from bottom of rad, up to CPU block and back to rad. Extra tube from res (right) was for easy drainage. 'T' in tube from CPU block to rad is not there now.
Proposed
1. Will reverse path so block is fed by rad and pump feeds rad from bottom.

2. Will add fill/bleed lines into top and bottom tanks (still figuring final placement). Top will be to bleed off (use top tank as airtrap since it seems to be that way now :mad: ) Bottom tank would be tapped for drainage. This would eliminate need for dedicated reservoir.

Don't know if that is needed, the Eheim 1250 is a good pump, not sure about the MCP 600. I would only consider the swap if the new pump had both higher head pressure numbers and better volumes.
Optimal flow path is Res => pump => rad => CPU => res. It is best if the water travels from the pump at lowest part of the system to the res at the highest. Other than when going from the res to the pump, plan your hose routing so that the water always goes uphill (ie flow enters rad at bottom and exits at top...) This will reduce trapped air to a minimum, as all the bubbles will end up in the res. If the res isn't at the high point, it MIGHT be worth adding some sort of bleeder at the high point.

IMPORTANT - Ensure there is little or no restriction in the line between the res and the pump inlet! You want the pump to be working on a positive or zero NPSH (Net Pump Suction Head) to avoid cavitation which causes noise and extra heat. Many folks make the line bigger than in the rest of the system so as to ensure as little flow restriction as possible.
T-lines work, but are less effective than res setups, and are harder to work with. Keep the res.

Gooserider

nice theory gooserider, one comment. if your flow is soo bad that it can't even take the bubbles down to a res that might be at the bottom, i advise to work on the flow some.

You are probably right, but I feel that a top mounted res (or one like mine that runs from the top of the case to the bottom) will give faster bubble elimination. I prefer working with natural forces rather than attempting to beat them with brute force...

Gooserider

*nods* A res at the top of the system will be best to eliminate air pockets - that space above the rad might be a good option - maybe a homemade res there would work nicely - you could even cut a window out and put a few lights behind and make a feature out of it :D.

~ Boli

You could, but I wouldn't... IMHO Windows are as problematic in a case as they are as an O/S on the hard drive. ;) While they let you see the inside of the case (IMHO not a big thrill...) they also let out lots of noise. I prefer my setup with a nice layer of asphalt roofers tape lining all exterior sheet metal, and a good bit of the interior metal to suppress resonant noise, followed up with a couple layers of carpet padding to absorb radiated sound... I'm probably not going to get to 'inaudible' (the SPCR gold standard) but I do want to approach 'extremely quiet'

Gooserider