Plumbing update....and thanks to all of you inquiring on the status of the build. Still been spending too much time playing Ingress...but starting to slow on that. Plus...I feel like I was kind of forced into an update based on my request for more FG pictures in another thread.
First, lets knock out some pictures of some small updates from last set of pictures. Both of these changes were suggested by members...and I think they improve the aesthetics just a little bit.
Previously, I had a small piece of crystal link coming out of one of the Q blocks that was the pathway from the GPU block to the board block. To help break-up the long run of white fittings, I moved the crystal link portion of the run to the middle of the path instead of having it start right after the Q block. Helps to break up the white and center the crystal link a little more. Improvement, I think.Old Layout (Click to show)
New Layout (Click to show)
Next, here is another small change that, again, is intended to break up one of the longer runs of white fittings with some crystal link. I had originally used a run of white fittings from one reservoir to the other. Someone suggested breaking that up with some crystal link...and I think this is an improvement also.Old Layout (Click to show)
New Layout (Click to show)
As a reminder, my first plumbing layout attempt here is to have two subloops (CPU subloop and a GPU subloop) that both share one radiator loop for maxium cooling efficiency. The CPU subloop order goes VR > RAM > CPU > RAM > AQ5 > AQ5. The GPU subloop order goes GPU1 > GPU2 > Main Board Block. Here is a full picture of the motherboard side showing these two subloops and the changes mentioned above:
However, most of the time I've spent on the plumbing since my last update, has been on the PSU side of the build. I'm starting with a variation of one of the many sketches that Electrocutor brainstormed for me. It uses 3x valves in the loop to control the flow between the CPU subloop and the GPU subloop. First, here are a couple of overall pictures, and then I'll try and explain it in more detail:
I've included some numbers and letters on the picture below to aid in describing what I'm trying to do. Follow the discussion while looking at the referenced #'s and letters.
(1) Flow from 2x pumps in main compartment
(2) Flow to CPU subloop
(3) Return flow from CPU subloop
(4) Flow to GPU subloop
(5) Flow to AQ5's and then ultimately the radiator loop.
The two pumps in the main compartment send all the flow through the fittings along path (1). The flow reaches a Y connector and this is where the first option exists. If valve (A) is closed, then all of the flow heads up the tubing path (2) to the CPU subloop. If valve (A) is open, then the flow is split between the CPU subloop and the GPU subloop. And of course, there are varying degrees of "open" in-between open and closed...which would allow me to control how much flow gets directed to the CPU subloop vs the GPU subloop.
The CPU subloop returns all the flow it received back to the PSU side of the build along tubing path (3). It reaches another Y connector, which creates another option for the flow at this point. If valve (C) is closed and valve (B) is open, then 100% of the flow goes to the GPU subloop via fitting path (4). If valve (C) is open and valve (B) is closed, then 100% of the flow goes to the AQ5's and radiator loop via tubing path (5). And again...there could be varying degrees of "open" which would allow fine tuning of how much flow goes to the GPU subloop vs the radiator loop.
Any flow directed to the GPU subloop ends up being directed to the radiator loop, but that happens on the other side of the box.
So consider the following scenarios, which I could easily switch between just by opening or closing valves. When I'm using the words serial and parallel below...I'm not referring to the parallel GPU blocks...I'm referring to the subloops. Whether or not the GPU blocks are in serial or parallel flow, has no bearing on the options below and could be changed independently:100% Serial Loop
-- Valves (A) and (C) closed, and valve (B) open. This would direct 100% of the flow to the CPU subloop, and then the GPU subloop, and then the radiator loop. It would be as-if there were not separate subloops at all.100% Parallel Subloops
-- Valves (A) and (C) open, and valve (B) closed. This would split the flow from the pumps immediately between the CPU subloop and the GPU subloop, and there would not be any intermingling of flow until it reached the radiator loop. Variations within this option could be controlled by how open valve (A) is. For example, if valve (A) were only 1/2 open, this would cause greater restriction in the GPU subloop and direct more of the initial flow to the CPU subloop.100% CPU Loop
-- Valves (A) and (B) closed, and valve (C) open. This would send all of the flow to the CPU subloop with zero flow going to GPU subloop. This might be useful if swapping video cards out and wanting to temporarily use an air cooler for GPU's. I could achieve this without having to make changes to the loop. Just opening or closing valves.100% GPU Loop
-- To accomplish this, I would really need to add a 4th valve on the other side of the first Y connector, and close it so that all of the flow goes toward tubing run (4). Valve (A) would be open, and valves (B) and (C) would be closed. Might add a 4th valve in a future update.
I'm excited about the thought of playing around with how flow impacts temperatures of the components in all these different options. I may find out that the impact is not large enough to excite me...and if so, I could probably make the tubing path less complicated. But for now...I like the options. I like the idea of being able to make changes on the fly just be adjusting valves. If I'm doing something CPU intensive...I could increase the flow to the CPU subloop. If I'm doing something GPU intensive....I could increase the flow to the GPU subloop. Options are good. Options are fun.
I think I'm ready to fill it up for a leak test. But before I can do that...I'll need to get it down the stairs to the kitchen. So perhaps not until this weekend. But very soon. I'm anxious to get to start using the beast daily!