please tell if this is a stupid idea

He may be right. I mite lose 2% flow rate which will equal to 0c-1c temp difference while cooling some parts that are going to allow for a higher oc ...I could go around pointing things like that out but that would be trolling...there are better ways to give advice.

streetsports said:

He may be right. I mite lose 2% flow rate which will equal to 0c-1c temp difference while cooling some parts that are going to allow for a higher oc ...I could go around pointing things like that out but that would be trolling...there are better ways to give advice.

Click to expand...

Well you should understand the position he's in when reading the title to this thread and putting 2 and 2 together when seeing a high restrictive loop in a pic. He know's his stuff as most of us do as we've been doing this stuff for quiet some time. Not here to argue, so we'll move along.

Not sure exactly how much flow will be lost to equate to higher temps but my guess is will be higher than 1c but as long as there is enough heat surface for it all it shouldn't matter much. Ram water cooling is highly restrictive and adding a lot of angle fitting etc will just add to that. It's not an issue as long as you have the pumping power. 1.0GPM-1.5GPM in general is the sweet spot. If you can get your constant flow into that range you should be good.

I understand what you're saying...you have a good way of giving advice...the ram block is the ek quad ram block...its not as restrictive as the koolance types. I used alot of angle fittings to keep the total loop as short as possible (it makes it so that 2/3 of the loop is radiator space...head pressure n whatnot). It also makes for a cleaner tube setup that won't block air flow as much as having a bunch of 1 foot long loops connecting everything...

My comment comes from what i perceive to be the order of the components...
there seems to be a lot of restriction BEFORE the CPU is hit.
And because of the CPU block impingement design with jetplate the flow/velocity might be low enough to affect performance is a negative way.

In other words, you might not get your moneys worth out of that otherwise excellent CPU block with this routing loop/setup.

I love a "clean" design as the next man, but i also like the best performance for the least effort; so i tend to take that into account.

But it all comes down to what you like and find acceptable.


.

I was thinking the same thing about the jet plate...I was going to remove it but it seems like it directs the water directly over the dies... I wonder if removing it would help

using martinsliquid lab estimator, which has always been close in estimating my setups with single or dual pumps compared to direct measurements, you are probably around 0.6 gpm mark with that setup (guessing some on ram blocks, though have seen restriction on some) (your setup is one that does benefit from parallel gpu).

Here at skinneelabs heatkiller testing, your cpu temp is probably about 2.5-3C higher at ~0.6gpm than it would be at 1.5gpm. jet plates typically only induce about 0.2-0.3 gpm restriction in typical loop but reduce temps by a few to several C. So removing the jet plate will likely increase temps (1C lower temps from increased flow but few to several C higher temps from loss of turbulence/directed flow).

A lot of people that go bling route with mobo blocks use a second pump and parallel gpus (you did parallel), you would probably get near 1 gpm with second pump or closer to 1.2 with dual pumps and aftermarket tops.

I'm using the mcp655 pump
Maximum head: 10 ft (3.1 m)
Maximum discharge: ~ 317 GPH (1200 LPH)
Connection size: 1/2" barbs (3/8" w/ conversion kit)
Maximum pressure: 50 PSI (3.5 BAR)

With the koolance top.

I changed my loop around so that it pushes through the 4 rads and then into the cpu 1st...getting ready to cut tube.

A second pump is now on the top of my list as well as a flow meter...thanks for the help. This is my first build.

loop order wont matter. It is a closed loop so although pressure differences will exist, they are simply added to the whole restriction of the loop, and flow through each component (what determines temps) will be same regardless of loop order.

With second pump with aftermarket tops, your flow should be fine.

That's what I thought...I only changed it because of what rnrollie said...thanks again.

rge said:

loop order wont matter. It is a closed loop so although pressure differences will exist, they are simply added to the whole restriction of the loop, and flow through each component (what determines temps) will be same regardless of loop order.

With second pump with aftermarket tops, your flow should be fine.

Click to expand...

+1... the only thing that matters in loop order is keeping a res before the pump to make sure it is primed.

This is going to look awesome! But what was said was right, cooling the ram and power bits on the board nets nothing performance but does have more restrictive parts (ram and mobo vrm blocks). It wont really matter in the end, but what is right is right.

Looking forward to seeing it done!

I updated the op with pics of the new setup...my temps are actually lower...2c Lowe on CPU, 9c lower on the mobo and 1c lower on the gpu's...

Edit.

22.7c ambient room temp
26c mobo! (This one kept dropping for an hour after first boot...it was running 38c idle before lol)
28c CPU idle
26c gpu's idle

I think its mostly from going from the r4 to the 900d with 19 fans

Here is where it sits after a couple hours at idle.